Networking Technologies

Because Ethernet LANs use contention for media access control, increase in the number of connected devices generally results in an increase in the number of collisions, eventually reducing network performance. To reduce collision problems, three common solutions can be implemented. First, bridges allow segments to be physically separated, reducing unnecessary traffic and congestion. Secondly, Ethernet LANs can be upgraded to 100 Mbps or ATM, etc. Thirdly, high-speed switching hubs or LAN switches can reduce congestion within a given workgroup or department.

Recently the release of 802.3u - Fast Ethernet, has defined standards by which networks can attain 100 Mbps and higher throughput rates. 802.3u includes standards for 100BASE-TX, 100BASE-T4 and 100BASE-FX. 100BASE-TX describes a dual-twisted pair implementation, while 100BASE-T4 describes four pairs of UTP wire implementations. Both allow for up to 100 Mbps throughput, using the same CSMA/CD media access methods, and with the same physical limitations on the media segment lengths. 100BASE-FX describes a fiber optic standard. 802.3u also describes two new classes of repeaters: Class I and Class II. Only one Class I repeaters may be cascaded in a collision domain, while only two Class II repeaters may be cascaded in a collision domain.

Because Ethernet LANs use contention for media access control, increase in the number of connected devices generally results in an increase in the number of collisions, eventually reducing network performance. To reduce collision problems, three common solutions can be implemented. First, bridges allow segments to be physically separated, reducing unnecessary traffic and congestion. Secondly, Ethernet LANs can be upgraded to 100 Mbps or ATM, etc. Thirdly, high-speed switching hubs or LAN switches can reduce congestion within a given workgroup or department. In the case of a switch, a single device or a hub may connect to a port of the switch, defining what is called a collision domain at that port, consisting of all devices that will contend for media access and bandwidth at that port. Switches are either cut-through or store-and-forward in their design. Cut-through switches does not read and verify each frame, but only reads the destination address prior to forwarding. While this greatly improves throughput at the switch, it also increases the potential for problems with bad frames, etc. Store-and-forward switches perform similarly to bridges, examining the entire frame before making the decision whether or not to forward.

Routers are used to interconnect networks, often located some distance from each other. To function effectively, a router needs the destination MAC address and a known path to that destination network. The addressing scheme used may be logical, where the address provided needs to be resolved (as in IP addressing), or a combination of physical and logical, where both are provided (as in IPX). In either case, the router must have information about known networks and paths to those networks, all referred to as routing tables, to effectively forward the packet to its eventual destination network. It is through network layer addressing that routers build these routing tables and determine the best path for forwarding. Protocols that do not define a network layer address are considered non-routable. To forward packets from these networks requires either bridging or encapsulation or tunneling.

The most common TCP/IP protocols are IP and ICMP (Internet layer), ARP, RARP, and BOOTP (address resolution), TCP and UDP (Host-to-Host layer) and the Process/Application layer protocols of FTP, TFTP, HTTP, SMTP, and SNMP. The address resolution protocols ARP, RARP and BOOTP all help resolve physical addresses and associated IP addresses. Since delivery of packets requires the destination device physical (MAC) address, ARP is used to map IP addresses to the associated MAC address. Devices maintain ARP tables cached generally in RAM, but can use broadcasts to help find the physical address if necessary. RARP provides the same service in the opposite direction. Used primarily for diskless workstations, RARP will "tell" them their IP address. Lastly, BOOTP is a newer version of RARP, utilizing a BOOTP server to maintain the lists of device MAC addresses and associated IP addresses. The most recent improvement on this process of providing IP addresses to hosts involves Dynamic Host Configuration Protocol (DHCP). With DHCP, the host issues a broadcast request for the nearest DHCP server, and when it hears a response, then asks for an address. DHCP addresses come in three forms: automatic, dynamic and manual. With automatic, DHCP assigns a permanent IP address to a given host. With dynamic, IP addresses are given to hosts for a specified period of time (referred to as a lease). With manual allocation, the administrator sets the IP address that the DHCP server will give a pre-determined host upon request.

WAN protocols allow an organization to connect remote locations. Generally these solutions involve leasing or purchasing throughput from a provider such as a PSTN (Public Switched Telephone Network). Protocols described for these solutions include PPP, SLIP, X.25, Frame Relay, ISDN and B-ISDN, and ATM.

OSPF is the link state routing protocol in IP. OSPF routers exchange hello packets upon initialization to determine neighbors and set the frequency for hello packet exchange, and then exchange them periodically to identify changes in the network topology. After this, OSPF routers only broadcast link state packets (LSPs), containing information about directly connected networks, rather than their entire table. Additionally, link state routers do not broadcast LSPs periodically, but only when they initialize, or when a change occurs in the internetwork. This reduces network traffic considerably. Because routers only send LSPs containing information about directly connected networks, other routers cannot receive "second hand" information. These routers will forward LSPs immediately, rather than first calculating their routing table. Thus convergence is much faster, and count-to-infinity does not occur. Once link state routers receive LSPs from all other routers, they then construct a "map" of the internetwork and from that build their routing table.

FDDI is very similar to IEEE 802.5, both use token-passing, both call for a star topology, both have many automated management features, and both can use fiber as the physical transmission media. FDDI differs in a number of ways: it can attain higher throughput rates, and so is often used for backbones and in other high-speed implementations. FDDI also uses a dual counter-rotating ring for redundancy. In this way, ring failure can be automatically routed back along the redundant ring to maintain connectivity. FDDI defines two types of stations: Class A and Class B. Class A stations connect to both rings and so allow the best fault-tolerant connectivity. Class B stations only connect to the primary ring and do not allow for automatic re-routing for fault tolerance.

Routers are used to interconnect networks, often located some distance from each other. To function effectively, a router needs the destination MAC address and a known path to that destination network. The addressing scheme used may be logical, where the address provided needs to be resolved (as in IP addressing), or a combination of physical and logical, where both are provided (as in IPX). In either case, the router must have information about known networks and paths to those networks, all referred to as routing tables, to effectively forward the packet to its eventual destination network.

Connectivity devices in networking are used to extend the LAN media, to help transport packets across networks with different media access methods and for communication between entities that use incompatible protocols. Most commonly, repeaters and bridges are used to extend the LAN media beyond its physical limitations. Repeaters do this by regenerating or amplifying the signal. Bridges accomplish this by filtering out frames from other segments and keeping frames destined for local devices on the local segment - thus reducing traffic on all segments. Routers are used to connect networks with different media access methods and gateways translate between incompatible protocols.

An IP address of a host consists of four bytes or octets divided into two parts, a network address and a node or station address. Each byte is a decimal number between 0 and 255. The bytes are separated by periods, and the value of the first byte determines the class of the IP address. A process called subnetting, which allows an organization to create many IP networks from a single address, can further extend IP addresses. Because of the depletion of registered IP addresses, there are only Class A addresses available. For organizations that need to network more than the 255 hosts that a Class C address will support, supernetting allows multiple Class C addresses to be grouped together into a single network. The Class C addresses must be contiguous, and the first address' third byte must be divisible by 2.

As a packet is forwarded from router to router through the internetwork, the determination of best path is calculated based on cost: in RIP internetworks either hops, the number of routers to the destination network, or in tick count, which is a pre-determined amount of time per tick (for instance, a tick in IPX is 1/18 of a second). In link state networks, the cost can be hops or costs at each interface. In either case, costs can be either automatically calculated or manually set. Manually configuring costs to pre-determine routes is known as load balancing or path splitting.

OSPF is the link state routing protocol in IP. OSPF routers exchange hello packets upon initialization to determine neighbors and set the frequency for hello packet exchange, and then exchange them periodically to identify changes in the network topology. The exchange of hello packets in an OSPF network causes the election of a designated router (DR) and backup designated router (BDR). After this, OSPF routers synchronize with (enter into full-neighbor state) the DR and BDR, and then exchange link state packets (LSPs) containing information about directly connected networks, rather than their entire table. This allows the DR and BDR to "collect" complete first-hand information about the network and "distribute" that information to every other router. In this way, all routers have exactly the same map of the internetwork. Additionally, link state routers do not broadcast LSPs periodically, but only when they initialize, or when a change occurs in the internetwork. This reduces network traffic considerably. These routers will forward LSPs immediately, rather than first calculating their routing table. Thus convergence is much faster, and count-to-infinity does not occur. Once link state routers receive LSPs from all other routers, they then construct a "map" of the internetwork and from that build their routing table.

WAN protocols allow an organization to connect remote locations. Generally these solutions involve leasing or purchasing throughput from a provider such as a PSTN (Public Switched Telephone Network). Protocols described for these solutions include PPP, SLIP, X.25, Frame Relay, ISDN and B-ISDN, and ATM. X.25 defined connectivity to a packet-switched network using DTE (Data Terminating Equipment) and DCE's (Data Circuit Equipment). The DTE and DCE establish and maintain virtual circuits for end-to-end error control, while flow the vendor whose network is used to transport the packets provides control. X.25 solutions map to the three lower layers of the OSI model: physical, data link and network. \\\\\\ Frame relay specifications define connectivity to high-speed digital networks. Because of the reliable nature of the digital network, there is less need for error checking and so frame relay provides higher data rates. Companies order lines based on CIR (Committed Information Rates) of 56 Kbps, T-1 or T-3. A CIR guarantees that the throughput will never fall below the ordered rate. Because frame relay is bursty, it is only suitable for data transmission. It is often used along with X.25 or ISDN implementations. ISDN defines standards for connecting both analog and digital networks to high-speed digital service provider's networks. ISDN lines offer multiple channels, called bit pipes that allow the customer to send voice on one pipe and data on the other. B-ISDN is an enhancement of ISDN that utilizes fiber optic media for greater throughput rates, available in multiples of 155 Mbps.

Connectivity devices in networking are used to extend the LAN media, to help transport packets across networks with different media access methods and for communication between entities that use incompatible protocols. Most commonly, repeaters and bridges are used to extend the LAN media beyond its physical limitations. Routers are used to connect networks with different media access methods and gateways translate between incompatible protocols.

An autonomous system in IP is a group of networks administered as one large network. It is accomplished using routing protocols, called interior gateway protocols (IGP) within the autonomous system, and exterior gateway protocols (EGP) to interconnect separate autonomous systems. In IP routing, RIP and OSPF can be used as interior gateway protocols.

WAN protocols allow an organization to connect remote locations. Generally these solutions involve leasing or purchasing throughput from a provider such as a PSTN (Public Switched Telephone Network). Protocols described for these solutions include PPP, SLIP, X.25, Frame Relay, ISDN and B-ISDN, and ATM. X.25 defined connectivity to a packet-switched network using DTE (Data Terminating Equipment) and DCE's (Data Circuit Equipment). The DTE and DCE establish and maintain virtual circuits for end-to-end error control, while flow the vendor whose network is used to transport the packets provides control. X.25 solutions map to the three lower layers of the OSI model: physical, data link and network. Frame relay specifications define connectivity to high-speed digital networks. Because of the reliable nature of the digital network, there is less need for error checking and so frame relay provides higher data rates. Companies order lines based on CIR (Committed Information Rates) of 56 Kbps, T-1 or T-3. A CIR guarantees that the throughput will never fall below the ordered rate. Because frame relay is bursty, it is only suitable for data transmission. It is often used along with X.25 or ISDN implementations.

There are three types of bridges commonly used in network implementations: transparent, source-routing and source-routing transparent. Transparent bridges (also called learning bridges) operate by reading the MAC address of devices connected to its ports as the frames are forwarded. This process creates a filtering database that can then be used by the bridge to determine whether or not to forward a frame based on its destination MAC address. The filtering database can also contain entries made manually by the manufacturer. Source-routing bridges, commonly found in token ring networks, do not use filtering databases, but rather forward frames based on information contained in the frame header. The information is created as the source workstation sends a hello packet to the target workstation. Eventually the hello packet reaches the target workstation, having had the source-routing bridges in transit add their route information to the packet. The packet is then returned to the source workstation, which adds the information to its routing database, and is used for further transmissions between the workstations.

If a bridge fails, connectivity is lost, and so redundant bridges are often placed throughout the network to reduce the risk of failure. Unfortunately, these redundant bridges create a problem known as "bridging loops". This condition is created when the redundant bridges begin looping a frame (or frames) back and forth, constantly modifying their own filtering database based on what they see as new and more correct information. Uncorrected, bridging loops will eventually consume the bandwidth on a network and render it unusable. This can be corrected by implementing the spanning tree protocol for bridges. Essentially, an election process determines a hierarchy of bridges, with only one of each pair of redundant bridges in forwarding mode (known as the designated bridge) while the other bridge (called the backup bridge) puts its ports into blocking state. This is accomplished as bridges exchange BPDUs (bridge packet data unit) containing the bridge ID. The bridge with the lowest bridge ID becomes the root bridge, which will then drive the selection of designated bridges throughout the network. Once elected, the root bridge transmits BPDUs from all ports at 2-second intervals. As the other bridges process these, a field in the BPDU called port cost is updated by adding the port cost (typically manually assigned) to the value already present in the field. Designated bridges are selected as redundant pairs of bridges examine each other's BPDUs. The bridge with the lowest port cost becomes the designated bridge, while the other bridge puts itself in blocking state.

If a bridge fails, connectivity is lost, and so redundant bridges are often placed throughout the network to reduce the risk of failure. Unfortunately, these redundant bridges create a problem known as "bridging loops". This condition is created when the redundant bridges begin looping a frame (or frames) back and forth, constantly modifying their own filtering database based on what they see as new and more correct information. Uncorrected, bridging loops will eventually consume the bandwidth on a network and render it unusable. This can be corrected by implementing the spanning tree protocol for bridges. Essentially, an election process determines a hierarchy of bridges, with only one of each pair of redundant bridges in forwarding mode (known as the designated bridge) while the other bridge (called the backup bridge) puts its ports into blocking state. This is accomplished as bridges exchange BPDUs (bridge packet data unit) containing the bridge ID. The bridge with the lowest bridge ID becomes the root bridge, which will then drive the selection of designated bridges throughout the network. Once elected, the root bridge transmits BPDUs from all ports at 2-second intervals. As the other bridges process these, a field in the BPDU called port cost is updated by adding the port cost (typically manually assigned) to the value already present in the field. Designated bridges are selected as redundant pairs of bridges examine each others BPDUs. The bridge with the lowest port cost becomes the designated bridge, while the other bridge puts itself in blocking state.

Once a domain name is registered, the name and its associated IP address are added to the appropriate DNS (Domain Name Service) server creating what is called a Standard Zone. The server to which the entry is added is called the master name server and contains what is referred to as the authoritative database. All updates must be made at the master name server. The authoritative database is then replicated to secondary or replica name servers so that the name resolution process occurs more efficiently. ISPs maintain replica name servers so that their clients' requests for name resolution can be resolved locally (and quickly). The process for updating the database on the replica name server is called a zone transfer.

As bridges exchange BPDUs (bridge packet data unit) containing the bridge ID, the root bridge is selected, which will then drive the selection of designated bridges throughout the network. Once elected, the root bridge transmits BPDUs from all ports at 2-second intervals. As the other bridges process these, a field in the BPDU called port cost is updated by adding the port cost (typically manually assigned) to the value already present in the field. Designated bridges are selected as redundant pairs of bridges examine each other's BPDUs. The bridge with the lowest port cost becomes the designated bridge, while the other bridge puts itself in blocking state. This backup bridge will transmit a topology change notification (TCN BPDU) if it does not see a frame forwarded by the designated bridge within a specified amount of time. This TCN BPDU travels to the root bridge, placing all bridges in blocking state and then the root bridge triggers another process for selection of designated bridges through port costs.

As bridges exchange BPDUs (bridge packet data unit) containing the bridge ID, the root bridge is selected, which will then drive the selection of designated bridges throughout the network. Once elected, the root bridge transmits BPDUs from all ports at 2-second intervals. As the other bridges process these, a field in the BPDU called port cost is updated by adding the port cost (typically manually assigned) to the value already present in the field. Designated bridges are selected as redundant pairs of bridges examine each other's BPDUs. The bridge with the lowest port cost becomes the designated bridge, while the other bridge puts itself in blocking state. This backup bridge will transmit a topology change notification (TCN BPDU) if it does not see a frame forwarded by the designated bridge within a specified amount of time. This TCN BPDU travels to the root bridge, placing all bridges in blocking state and then the root bridge triggers another process for selection of designated bridges through port costs.

Connectivity devices in networking are used to extend the LAN media, to help transport packets across networks with different media access methods and for communication between entities that use incompatible protocols. Most commonly, repeaters and bridges are used to extend the LAN media beyond its physical limitations. Repeaters do this by regenerating or amplifying the signal. Bridges accomplish this by filtering out frames from other segments and keeping frames destined for local devices on the local segment - thus reducing traffic on all segments. Routers are used to connect networks with different media access methods and gateways translate between incompatible protocols.

Connectivity devices in networking are used to extend the LAN media, to help transport packets across networks with different media access methods and for communication between entities that use incompatible protocols. Most commonly, repeaters and bridges are used to extend the LAN media beyond its physical limitations. Repeaters do this by regenerating or amplifying the signal. Bridges accomplish this by filtering out frames from other segments and keeping frames destined for local devices on the local segment - thus reducing traffic on all segments. Bridged networks are designed based on the 80-20 rule which states that 80% of the traffic on a segment should be local, with no more than 20% intended to cross the bridge. There are three types of bridges commonly used in network implementations: transparent, source-routing and source-routing transparent. Transparent bridges (also called learning bridges) operate by reading the MAC address of devices connected to its ports as the frames are forwarded. This process creates a filtering database that can then be used by the bridge to determine whether or not to forward a frame based on its destination MAC address. The filtering database can also contain entries made manually by the manufacturer.

X.500 is a specification for directory services and calls for the maintenance of a database containing all of the directory information called the Directory Information Base or DIB. This database is a hierarchical structure with leaf objects representing directory entries and intermediate objects used to organize the leaf objects. Access to the DIB is requested using a Directory User Agent or DUA at the client side, and is fulfilled by a Directory Service Agent or DSA at the server side. Current examples of directories created based on the X.500 specifications include DNS, Novell's NDS and Microsoft's Active Directory.

WAN protocols allow an organization to connect remote locations. Generally these solutions involve leasing or purchasing throughput from a provider such as a PSTN (Public Switched Telephone Network). Protocols described for these solutions include PPP, SLIP, X.25, Frame Relay, ISDN and B-ISDN, and ATM. The earliest of these was SLIP, originally designed to provide dial-up connectivity for TCP/IP networks. SLIP implementations were simple, but lacked features and implementations from different organizations would often not work together. PPP was defined as a stricter standard, and offered many features not available in SLIP, for example, password security, dynamic IP addressing, error control and multiple protocol support.

WAN protocols allow an organization to connect remote locations. Generally these solutions involve leasing or purchasing throughput from a provider such as a PSTN (Public Switched Telephone Network). Protocols described for these solutions include PPP, SLIP, X.25, Frame Relay, ISDN and B-ISDN, and ATM. X.25 defined connectivity to a packet-switched network using DTE (Data Terminating Equipment) and DCE's (Data Circuit Equipment). The DTE and DCE establish and maintain virtual circuits for end-to-end error control, while flow the vendor whose network is used to transport the packets provides control. X.25 solutions map to the three lower layers of the OSI model: physical, data link and network. Frame relay specifications define connectivity to high-speed digital networks. Because of the reliable nature of the digital network, there is less need for error checking and so frame relay provides higher data rates. Companies order lines based on CIR (Committed Information Rates) of 56 Kbps, T-1 or T-3. A CIR guarantees that the throughput will never fall below the ordered rate. Because frame relay is bursty, it is only suitable for data transmission. It is often used along with X.25 or ISDN implementations. ISDN defines standards for connecting both analog and digital networks to high-speed digital service provider's networks. ISDN lines offer multiple channels, called bit pipes that allow the customer to send voice on one pipe and data on the other.

WAN protocols allow an organization to connect remote locations. Generally these solutions involve leasing or purchasing throughput from a provider such as a PSTN (Public Switched Telephone Network). Protocols described for these solutions include PPP, SLIP, X.25, Frame Relay, ISDN and B-ISDN, and ATM. Of these, ATM is unusual in that it is used in WAN and LAN environments. ATM is defined at the Network and Data Link layers, independent from the Physical layer. Often implemented along with SONET or FDDI, ATM creates a "virtual channel" using a series of pointers (generally hardware switches), along which "cells" are transported. ATM transmits data in cells (53 byte packets) with extremely low overhead (only 5 bytes), and in combination with fiber optic physical layer implementations, can reach throughput rates in excess of 155 and 622 Gbps. A group of standards called the I Series defines classes of service available in ATM implementations.

OSPF is the link state routing protocol in IP. OSPF routers exchange hello packets upon initialization to determine neighbors and set the frequency for hello packet exchange, and then exchange them periodically to identify changes in the network topology. The exchange of hello packets in an OSPF network causes the election of a designated router (DR) and backup designated router (BDR). After this, OSPF routers synchronize with (enter into full-neighbor state) the DR and BDR, and then exchange link state packets (LSPs) containing information about directly connected networks, rather than their entire table. This allows the DR and BDR to "collect" complete first-hand information about the network and "distribute" that information to every other router. In this way, all routers have exactly the same map of the internetwork. In the implementation of OSPF with autonomous systems, the backbone is always addressed as 0.0.0.0.

Connectivity devices in networking are used to extend the LAN media, to help transport packets across networks with different media access methods and for communication between entities that use incompatible protocols. Most commonly, repeaters and bridges are used to extend the LAN media beyond its physical limitations. Repeaters do this by regenerating or amplifying the signal. Bridges accomplish this by filtering out frames from other segments and keeping frames destined for local devices on the local segment - thus reducing traffic on all segments. Routers are used to connect networks with different media access methods and gateways translate between incompatible protocols. Bridges function at the data link layer of the OSI model.

Ring topologies involve a closed loop of point-to-point connections. Devices either connect directly to the ring or indirectly using an interface card and drop cable. IBM token ring uses hubs arranged in a physical ring, while user devices connect to the hubs in a star topology. IEEE 802.5 was taken directly from IBM's Token Ring specifications. Both specifications share some commonalities: each segment must be terminated at both ends and grounded, no more than three segments connected by repeaters, all devices must operate at the same speed (either 4 or 16 Mbps). With IEEE 802.5 the maximum number of connected nodes is 250, with IBM it is 260. In both cases the maximum number of MAU's is 33. In both specifications, automated troubleshooting functions are defined. While transmitting stations are responsible for removal of tokens after transmission, each ring has a device that is designated the active monitor and is responsible for removal of erroneous tokens as well as other administrative functions. In the event of a segment or device failure, the stations detecting the failure begin "beaconing", a process that aids in reconfiguration around the fault, and in detection of the fault.

WAN protocols allow an organization to connect remote locations. Generally these solutions involve leasing or purchasing throughput from a provider such as a PSTN (Public Switched Telephone Network). Protocols described for these solutions include PPP, SLIP, X.25, Frame Relay, ISDN and B-ISDN, and ATM. X.25 defined connectivity to a packet-switched network using DTE (Data Terminating Equipment) and DCE's (Data Circuit Equipment). The DTE and DCE establish and maintain virtual circuits for end-to-end error control, while flow the vendor whose network is used to transport the packets provides control. X.25 solutions map to the three lower layers of the OSI model: physical, data link and network. Frame relay specifications define connectivity to high-speed digital networks. Because of the reliable nature of the digital network, there is less need for error checking and so frame relay provides higher data rates. Companies order lines based on CIR (Committed Information Rates) of 56 Kbps, T-1 or T-3. A CIR guarantees that the throughput will never fall below the ordered rate. Because frame relay is bursty, it is only suitable for data transmission. It is often used along with X.25 or ISDN implementations. ISDN defines standards for connecting both analog and digital networks to high-speed digital service provider's networks. ISDN lines offer multiple channels, called bit pipes that allow the customer to send voice on one pipe and data on the other.

IPX addressing occurs at four different levels: network address (sometimes called external network address), internal network address, and node or MAC address and socket number. The network address is an 8-digit hexadecimal number that identifies the logical grouping of devices we refer to as a network. Two devices must be on the same network number to communicate with IPX. When IPX is bound to a network board, the IPX network number is either required to be manually provided or is automatically detected by the operating system. The internal network number is an 8-digit hexadecimal number that allows a server to internally route requests to the appropriate upper layer service. When you install NetWare, you assign an internal IPX network number. Servers cannot have the same internal network number. The third type of address is the node or MAC address, a 12-digit hexadecimal number generally hard-coded by the manufacturer of the NIC card. All devices on the same IPX network must have unique MAC addresses. The last type of address is the socket address, and specifies the port or socket of the entity within a device that provides the necessary service.

The application layer of the OSI model describes five methods for the advertisement of network services, two methods to use those services, and three methods for service use. The three service use methods are collaborative, remote operation and OS call interception. With remote operation the server is unaware of the client, with OS call interception the client is unaware of the server, and with collaborative both are aware.

Protocols in IPX include MLID and LSL at the data link layer, IPX, RIP and NLSP at the network and transport layers, SPX at the transport layer and NCP and SAP at the upper layers of the OSI model. NCP or NetWare Core Protocol functionality involves both client and server-side components. The NCP agents work as function calls using what are called well-known port or socket numbers for the IPX service being requested.

The most common TCP/IP protocols are IP and ICMP (Internet layer), ARP, RARP, and BOOTP (address resolution), TCP and UDP (Host-to-Host layer) and the Process/Application layer protocols of FTP, TFTP, HTTP, SMTP, and SNMP. At the Internet layer, IP provides connectionless, non-guaranteed delivery of packets while supporting routing, fragmentation and reassembly of packets. ICMP works with IP to help hosts determine network conditions, for example, congestion. The Ping command is an implementation of ICMP.

WAN protocols allow an organization to connect remote locations. Generally these solutions involve leasing or purchasing throughput from a provider such as a PSTN (Public Switched Telephone Network). Protocols described for these solutions include PPP, SLIP, X.25, Frame Relay, ISDN and B-ISDN, and ATM. The earliest of these was SLIP, originally designed to provide dial-up connectivity for TCP/IP networks. SLIP implementations were simple, but lacked features and implementations from different organizations would often not work together. PPP was defined as a stricter standard, and offered many features not available in SLIP, for example, password security, dynamic IP addressing, error control and multiple protocol support.

OSPF is the link state routing protocol in IP. OSPF routers exchange hello packets upon initialization to determine neighbors and set the frequency for hello packet exchange, and then exchange them periodically to identify changes in the network topology. After this, OSPF routers only broadcast link state packets (LSPs), containing information about directly connected networks, rather than their entire table. Additionally, link state routers do not broadcast LSPs periodically, but only when they initialize, or when a change occurs in the internetwork. This reduces network traffic considerably. Because routers only send LSPs containing information about directly connected networks, other routers cannot receive "second hand" information. These routers will forward LSPs immediately, rather than first calculating their routing table. Thus convergence is much faster, and count-to-infinity does not occur. Once link state routers receive LSPs from all other routers, they then construct a "map" of the internetwork and from that build their routing table.

WAN protocols allow an organization to connect remote locations. Generally these solutions involve leasing or purchasing throughput from a provider such as a PSTN (Public Switched Telephone Network). Protocols described for these solutions include PPP, SLIP, X.25, Frame Relay, ISDN and B-ISDN, and ATM. X.25 defined connectivity to a packet-switched network using DTE (Data Terminating Equipment) and DCE's (Data Circuit Equipment). The DTE and DCE establish and maintain virtual circuits for end-to-end error control, while flow the vendor whose network is used to transport the packets provides control. X.25 solutions map to the three lower layers of the OSI model: physical, data link and network.

IEEE 802.3 was based on the original Ethernet specifications, and evolved into numerous different implementations, all using CSMA/CD (collision detection) for media access, but each describing different media type, signaling methods, topologies and throughput rates. The most common are the 10BASE series of protocols, of which 10BASET is the most widely known and implemented. It calls for UTP cabling with a hierarchical star topology. Today's LAN environments are commonly cabled this way using CAT5 cable, hubs and switches.

X.500 is a specification for directory services and calls for the maintenance of a database containing all of the directory information called the Directory Information Base or DIB. This database is a hierarchical structure with leaf objects representing directory entries and intermediate objects used to organize the leaf objects. Access to the DIB is requested using a Directory User Agent or DUA at the client side, and is fulfilled by a Directory Service Agent or DSA at the server side. Current examples of directories created based on the X.500 specifications include DNS, Novell's NDS and Microsoft's Active Directory.

Connectionless protocols will provide services between entities without exchanging setup messages. In most LAN protocol implementations, the data link and network layers are connectionless, with connection-oriented services provided by transport layer protocols.

NLSP is the link state routing protocol in IPX. NLSP routers exchange hello packets upon initialization to determine neighbors and set the frequency for hello packet exchange, and then exchange them periodically to identify changes in the network topology. The exchange of hello packets in an NLSP network causes the election of a designated router (DR) and backup designated router (BDR). The router with the highest priority number becomes the DR. After this, the Designated Router creates what is called a pseudonode, and all other routers see themselves as directly connected to this virtual node. This effectively reduces the size of any given router's LSP (created from the adjacency database) and streamlines the entire process. Then the routers exchange LSPs to begin the process of building the link state database.

Additionally, link state routers do not broadcast LSPs periodically, but only when they initialize, or when a change occurs in the internetwork. This reduces network traffic considerably. These routers will forward LSPs immediately, rather than first calculating their routing table. Thus convergence is much faster, and count-to-infinity does not occur. Once link state routers receive LSPs from all other routers, they then construct a "map" of the internetwork and from that build their routing table.

The application layer of the OSI model describes five methods for the advertisement of network services, two methods to use those services, and three methods for service use. The three service use methods are collaborative, remote operation and OS call interception. With remote operation the server is unaware of the client, with OS call interception the client is unaware of the server, and with collaborative both are aware.

All networks require three elements: individuals with something to share, a pathway to communicate and rules for communication. In modern networks, services are capabilities that computers share. Combining these services into a single entity has given rise to the creation of network operating systems (NOS). A NOS generally will run on a designated computer and provide services to clients that communicate requests to the server. The most common services include file, print, message, database and application services. Database services provide server-based control of distributed, replicated databases. These services include control in terms of where data is stored, tuning in terms of where processing is performed, security, etc. To the users and administrators, the database may seem to be a single, seamless entity, while in fact no single server contains the entire database. This is achieved through the distribution of replicas, or copies, of portions of the database. These replicas can be placed near users or administrators to improve access time and reduce network traffic, while also improving fault tolerance for the database. NDS, Novell Directory Services, DNS< Domain Name Services, and Active Directory are all commercial examples of implementations of database services.

In computer networks, electric currents, radio waves, microwaves or light photons are used for encoding and decoding of data transmissions. These transmission methods require a physical pathway to carry them - this physical pathway we refer to as the transmission media, and can be wire or wireless in physical form. Transmission media has characteristics that can be used for comparison: cost and ease of installation, capacity, attenuation and immunity from interference and signal capture.

Cost and ease of installation is often determined by the availability of transmission solutions, the complexity of the necessary equipment, governmental regulation (especially in wireless media), etc. Capacity, usually stated in terms of bandwidth, is a measure of how much binary data a transmission media can reliably carry. We also use the term throughput to describe capacity. Attenuation is a term used to describe the tendency of electromagnetic signals to weaken or become distorted as they travel through transmission media. Signal interference occurs when outside electromagnetic waves mix with the transmission and often comes form two sources: EMI (Electromagnetic Interference) generated by large motors in industrial equipment, and RFI (Radio Frequency Interference) created by radio transmitters. Signal capture is the interception of the transmission, and is generally considered a security issue. Immunity from these factors is the fourth consideration when evaluating transmission media.

All networks require three elements: individuals with something to share, a pathway to communicate and rules for communication. In modern networks, services are capabilities that computers share. This has given rise to classifying of networks as either peer-to-peer networks, where each computer may request and provide services, or server-centric networks, where roles are more strictly defined; that is, in server-centric networks we generally restrict which computers can request and which computers provide services. We then call computers that provide services servers, and we call computers that request services clients. The second element, a pathway to communicate, includes both wire and wireless transmission media. The third element, rules to communicate, has given rise to protocols.

In computer networks, electric currents, radio waves, microwaves or light photons are used for encoding and decoding of data transmissions. These transmission methods require a physical pathway to carry them - this physical pathway we refer to as the transmission media, and can be wire or wireless in physical form. Transmission media has characteristics that can be used for comparison: cost and ease of installation, capacity, attenuation and immunity from interference and signal capture. Signal interference occurs when outside electromagnetic waves mix with the transmission and often comes form two sources: EMI (Electromagnetic Interference) generated by large motors in industrial equipment, and RFI (Radio Frequency Interference) created by radio transmitters. Signal capture is the interception of the transmission, and is generally considered a security issue. Immunity from these factors is the fourth consideration when evaluating transmission media.

All networks require three elements: individuals with something to share, a pathway to communicate and rules for communication. In modern networks, services are capabilities that computers share. The second element, a pathway to communicate, includes both wire and wireless transmission media. The third element, rules to communicate, has given rise to protocols. In computer networks, electric currents, radio waves, microwaves or light photons are used for encoding and decoding of data transmissions. These transmission methods require a physical pathway to carry them - this physical pathway we refer to as the transmission media, and can be wire or wireless in physical form.

In computer networks, electric currents, radio waves, microwaves or light photons are used for encoding and decoding of data transmissions. These transmission methods require a physical pathway to carry them - this physical pathway we refer to as the transmission media, and can be wire or wireless in physical form. Wire transmission media in computer networks conduct either electrical signals or lights. There are three commonly used types of cable media: twisted pair cable, coaxial cable and fiber optic cable. Twisted pair cable consists of insulated 22 to 26 gauge copper wires twisted around each other. The twisting reduces crosstalk and helps cancel signal emissions from the individual wires. Twisted pair cable comes in shielded and unshielded, with the shielded providing better immunity from EMI.

All networks require three elements: individuals with something to share, a pathway to communicate and rules for communication. In modern networks, services are capabilities that computers share. This has given rise to classifying of networks as either peer-to-peer networks, where each computer may request and provide services, or server-centric networks, where roles are more strictly defined; that is, in server-centric networks we generally restrict which computers can request and which computers provide services. We then call computers that provide services servers, and we call computers that request services clients. The second element, a pathway to communicate, includes both wire and wireless transmission media. The third element, rules to communicate, has given rise to protocols.

All networks require three elements: individuals with something to share, a pathway to communicate and rules for communication. In modern networks, services are capabilities that computers share. Combining these services into a single entity has given rise to the creation of network operating systems (NOS). A NOS generally will run on a designated computer and provide services to clients that communicate requests to the server. The most common services include file, print, message, database and application services. File services help clients to store, retrieve or move data. File services normally include file transfer, file storage and data migration, and file update synchronization services.

In computer networks, electric currents, radio waves, microwaves or light photons are used for encoding and decoding of data transmissions. These transmission methods require a physical pathway to carry them - this physical pathway we refer to as the transmission media, and can be wire or wireless in physical form. Wire transmission media in computer networks conduct either electrical signals or lights. There are three commonly used types of cable media: twisted pair cable, coaxial cable and fiber optic cable. Twisted pair cable consists of insulated 22 to 26 gauge copper wires twisted around each other. The twisting reduces crosstalk and helps cancel signal emissions from the individual wires. Twisted pair cable comes in shielded and unshielded, with the shielded providing better immunity from EMI. Shielded twisted pair does not improve attenuation, but achieves higher throughput rates due to its better immunity from interference.

Wireless media are generally categorized as radio wave, microwave or infrared. Radio wave transmission, between 10 KHz and 1 GHz, divides between regulated and unregulated bands. Use of regulated frequencies can guarantee transmission, but licensing is costly and time-consuming. The use of unregulated frequencies generally has limitations in terms of the power levels of the devices. This then limits the effective range for transmission. Radio wave systems include low power, single-frequency, high power single-frequency, and spread spectrum. Low power, single frequency systems are relatively inexpensive, usually easy to install, have high attenuation and are very susceptible to interference. These systems are used in short-distance, open environments. Garage door openers are such devices.

High power single-frequency systems are a bit more costly than their low power cousins, but installation is often much more complex, because of the precise tuning required and the high voltages used. These systems attenuate much slower than low power systems, are also very susceptible to interference, and are very prone to signal capture. Spread spectrum uses two different modulation schemes to resist signal capture: direct sequence modulation and frequency hopping. Direct sequence modulation spreads data (chips) across a subset of radio frequencies. Frequency hopping switches between predetermined frequencies, requiring both the sender and receiver to follow complex timing patterns to maintain communications. These solutions usually are low power, operating in unregulated frequencies, are moderately expensive pre-configured systems, suffer from high attenuation, and are somewhat susceptible to interference.

In computer networks, electric currents, radio waves, microwaves or light photons are used for encoding and decoding of data transmissions. These transmission methods require a physical pathway to carry them - this physical pathway we refer to as the transmission media, and can be wire or wireless in physical form. Wire transmission media in computer networks conduct either electrical signals or lights. There are three commonly used types of cable media: twisted pair cable, coaxial cable and fiber optic cable. Coaxial cable has two conductors that share a common axis. Commonly used in cable television, it is no longer in favor in PC networks. It is installed using a linear bus topology; with devices connected to a main trunk using smaller segments of coax called drops or drop cables. The two ends of the trunk must be terminated to remove the signal from the wire before it can reflect back along the pathway and the cable must be grounded to remove voltage from the shield. Because of the thicker outer shielding and the larger center conductor, coax offers better immunity from interference, and offers higher capacity than twisted pair.

In computer networks, electric currents, radio waves, microwaves or light photons are used for encoding and decoding of data transmissions. These transmission methods require a physical pathway to carry them - this physical pathway we refer to as the transmission media, and can be wire or wireless in physical form. Wire transmission media in computer networks conduct either electrical signals or lights. There are three commonly used types of cable media: twisted pair cable, coaxial cable and fiber optic cable. Twisted pair cable consists of insulated 22 to 26 gauge copper wires twisted around each other. The twisting reduces crosstalk and helps cancel signal emissions from the individual wires. Twisted pair cable comes in shielded and unshielded, with the shielded providing better immunity from EMI. Unshielded twisted pair cable (UTP) generally comes in either Category 3 or Category 5. The higher the category, the higher the data transmission rate, often due to more exacting installation requirements and the use of higher grade insulators and using more twists per foot.

Wireless media are generally categorized as radio wave, microwave or infrared. Radio wave transmission, between 10 KHz and 1 GHz, divides between regulated and unregulated bands. Use of regulated frequencies can guarantee transmission, but licensing is costly and time-consuming. The use of unregulated frequencies generally has limitations in terms of the power levels of the devices. This then limits the effective range for transmission. Radio wave systems include low power, single-frequency, high power single-frequency, and spread spectrum. Low power, single frequency systems are relatively inexpensive, usually easy to install, have high attenuation and are very susceptible to interference. These systems are used in short-distance, open environments. Garage door openers are such devices.
High power single-frequency systems are a bit more costly than their low power cousins, but installation is often much more complex, because of the precise tuning required and the high voltages used. These systems attenuate much slower than low power systems, are also very susceptible to interference, and are very prone to signal capture. Spread spectrum uses two different modulation schemes to resist signal capture: direct sequence modulation and frequency hopping. Direct sequence modulation spreads data (chips) across a subset of radio frequencies. Frequency hopping switches between predetermined frequencies, requiring both the sender and receiver to follow complex timing patterns to maintain communications. These solutions usually are low power, operating in unregulated frequencies, are moderately expensive pre-configured systems, suffer from high attenuation, and are somewhat susceptible to interference.

Coaxial cable has two conductors that share a common axis. Commonly used in cable television, it is no longer in favor in PC networks. It is installed using a linear bus topology; with devices connected to a main trunk using smaller segments of coax called drops or drop cables. The two ends of the trunk must be terminated to remove the signal from the wire before it can reflect back along the pathway and the cable must be grounded to remove voltage from the shield. Because of the thicker outer shielding and the larger center conductor, coax offers better immunity from interference, and offers higher capacity than twisted pair.

Wireless media are generally categorized as radio wave, microwave or infrared. Radio wave transmission, between 10 KHz and 1 GHz, divides between regulated and unregulated bands. Use of regulated frequencies can guarantee transmission, but licensing is costly and time-consuming. The use of unregulated frequencies generally has limitations in terms of the power levels of the devices. This then limits the effective range for transmission. Radio wave systems include low power, single-frequency, high power single-frequency, and spread spectrum.
Microwave systems exist in either terrestrial or satellite versions. Terrestrial systems use line-of-sight parabolic antennas, pre-configured equipment and often require both licensing for the frequencies used and complex installation techniques. Satellite microwave extends the path to satellite transceivers parked in a geosynchronous orbit above the earth. Both are capable of high rates of transmission. Satellite solutions are prone to delays in transmission because of the distances involved, called propagation delays.

Wire transmission media in computer networks conduct either electrical signals or lights. There are three commonly used types of cable media: twisted pair cable, coaxial cable and fiber optic cable. Fiber optic cabling conducts light waves along optical fibers. Generally used for high-speed data transmission, fiber optic is expensive because of the exacting requirements for the manufacture of the cable and the equipment used to connect devices. Fiber is unaffected by EMI or RFI, since it transmits using light pulses, and because it uses light for transmission rather than electrical currents, it is also less susceptible to attenuation.

Devices need to be able to connect to the transmission media, and the cable segments need to be connected together to allow for communication to proceed. Devices that are used to connect to media segments include transmission media connectors, network interface boards and modems. Devices that interconnect separate segments of transmission media include repeaters, hubs, bridges and multiplexors. Repeaters are used to amplify or regenerate the signal, thereby extending the media's maximum effective length. Repeaters are either amplifiers, which only boost the signal and will also amplify noise, or signal regenerating amplifiers that strips the data from the signal and then re-creates the transmission. Signal regeneration results in cleaner transmissions, but requires more complex equipment and introduces some delay due to the time needed to process the signal.
Hubs are used as a central point of connection for multiple segments of media, and include active, passive, multiport repeaters and switches. Bridges are used to extend the network by connecting separate network segments. Bridges only forward signals along ports if the signal is not intended for any device on the incoming port. They accomplish this by reading the physical device address (MAC address) of the packet, and so can process very quickly providing fast throughput. Because they selectively forward packets, bridges are often used to reduce traffic between segments, thereby reducing network congestion. Multiplexors combine the signals from multiple devices for transmission on a media segment and are used to allow for maximum use of the media's capacity.
Routers, brouters and CSU/DSU devices are used to connect multiple independent networks. Because these networks are independent, they are then identified by logical network addresses, forming what are then referred to as subnetworks or subnets. These devices require processing capabilities to identify the network address information inside an incoming packet and determine along which port to forward the packet for eventual delivery to the destination network. These devices generally are slower than their simpler network connectivity cousins, such as bridges, switches, etc.

In computer networks, electric currents, radio waves, microwaves or light photons are used for encoding and decoding of data transmissions. These transmission methods require a physical pathway to carry them - this physical pathway we refer to as the transmission media, and can be wire or wireless in physical form. Wire transmission media in computer networks conduct either electrical signals or lights. There are three commonly used types of cable media: twisted pair cable, coaxial cable and fiber optic cable.
Twisted pair cable consists of insulated 22 to 26 gauge copper wires twisted around each other. The twisting reduces crosstalk and helps cancel signal emissions from the individual wires. Twisted pair cable comes in shielded and unshielded, with the shielded providing better immunity from EMI. Coaxial cable has two conductors that share a common axis. Commonly used in cable television, it is no longer in favor in PC networks. Fiber optic cabling conducts light waves along optical fibers. Generally used for high-speed data transmission, fiber optic is expensive but is unaffected by EMI or RFI.

Devices that interconnect separate segments of transmission media include repeaters, hubs, bridges and multiplexors. Repeaters are used to amplify or regenerate the signal, thereby extending the media's maximum effective length. Repeaters are either amplifiers, which only boost the signal and will also amplify noise, or signal regenerating amplifiers that strips the data from the signal and then re-creates the transmission. Signal regeneration results in cleaner transmissions, but requires more complex equipment and introduces some delay due to the time needed to process the signal.

Devices that are used to connect to media segments include transmission media connectors, network interface boards and modems. Transmission media connectors are attached directly to the media, and allow the computing devices to connect to the media using network interface boards, which in our computer networks is usually a board that is installed into the computer to complete the connection between the computer and the media. Network interface boards are also referred to as transceivers, network interface cards (NIC) or transmission media adaptors. Modems are used when computers need to connect to analog networks. Modems convert the computer's digital transmission to analog for transmission, and then convert the incoming analog transmission to digital for the computer.

In computer networks, electric currents, radio waves, microwaves or light photons are used for encoding and decoding of data transmissions. These transmission methods require a physical pathway to carry them - this physical pathway we refer to as the transmission media, and can be wire or wireless in physical form. Wire transmission media in computer networks conduct either electrical signals or lights. There are three commonly used types of cable media: twisted pair cable, coaxial cable and fiber optic cable. Twisted pair cable consists of insulated 22 to 26 gauge copper wires twisted around each other. The twisting reduces crosstalk and helps cancel signal emissions from the individual wires. Twisted pair cable comes in shielded and unshielded, with the shielded providing better immunity from EMI.
Unshielded twisted pair cable (UTP) generally comes in either Category 3 or Category 5. The higher the category, the higher the data transmission rate, often due to more exacting installation requirements and the use of higher grade insulators and using more twists per foot. Coaxial cable has two conductors that share a common axis. Commonly used in cable television, it is no longer in favor in PC networks. It is installed using a linear bus topology; with devices connected to a main trunk using smaller segments of coax called drops or drop cables. The two ends of the trunk must be terminated to remove the signal from the wire before it can reflect back along the pathway and the cable must be grounded to remove voltage from the shield. Because of the thicker outer shielding and the larger center conductor, coax offers better immunity from interference, and offers higher capacity than twisted pair.
The most commonly used types of coaxial cable are 50 ohm RG-58 (thin Ethernet), used in computer networks to connect user devices to the network, 50 ohm RG-8/RG-11 (thick Ethernet), used in computer networks as the "backbone" that user devices connect to for transmission, 75-ohm RG-59, used for cable TV, and 93-ohm RG-62, used for ARC net.

Devices need to be able to connect to the transmission media, and the cable segments need to be connected together to allow for communication to proceed. Devices that are used to connect to media segments include transmission media connectors, network interface boards and modems. Devices that interconnect separate segments of transmission media include repeaters, hubs, bridges and multiplexors. Repeaters are used to amplify or regenerate the signal, thereby extending the media's maximum effective length. Repeaters are either amplifiers, which only boost the signal and will also amplify noise, or signal regenerating amplifiers that strips the data from the signal and then re-creates the transmission. Signal regeneration results in cleaner transmissions, but requires more complex equipment and introduces some delay due to the time needed to process the signal.
Hubs are used as a central point of connection for multiple segments of media, and include active, passive, multiport repeaters and switches. Bridges are used to extend the network by connecting separate network segments. Bridges only forward signals along ports if the signal is not intended for any device on the incoming port. They accomplish this by reading the physical device address (MAC address) of the packet, and so can process very quickly providing fast throughput. Because they selectively forward packets, bridges are often used to reduce traffic between segments, thereby reducing network congestion. Multiplexors combine the signals from multiple devices for transmission on a media segment and are used to allow for maximum use of the media's capacity.

Networking is the sharing of information and services. Within that definition, describe three basic models of computing. Collaborative computing involves two or more computer working on the same application, distributed involves clients and servers where the clients do processing of their own, and centralized computing is characterized by the use of mainframe computers that perform all of the processing and terminals that simply exchange screen and keyboard/mouse input. Client/server describes a particular implementation of distributed computing where several PC's are connected to a server (PC).

Networking is the sharing of information and services. Within that definition, describe three basic models of computing. Collaborative computing involves two or more computer working on the same application, distributed involves clients and servers where the clients do processing of their own, and centralized computing is characterized by the use of mainframe computers that perform all of the processing and terminals that simply exchange screen and keyboard/mouse input. Client/server describes a particular implementation of distributed computing where several PC's are connected to a server (PC).

Networking is the sharing of information and services. Within that definition, describe three basic models of computing. Collaborative computing involves two or more computer working on the same application, distributed involves clients and servers where the clients do processing of their own, and centralized computing is characterized by the use of mainframe computers that perform all of the processing and terminals that simply exchange screen and keyboard/mouse input. Client/server describes a particular implementation of distributed computing where several PC's are connected to a server (PC).

Networking is the sharing of information and services. Within that definition, describe three basic models of computing. Collaborative computing involves two or more computer working on the same application, distributed involves clients and servers where the clients do processing of their own, and centralized computing is characterized by the use of mainframe computers that perform all of the processing and terminals that simply exchange screen and keyboard/mouse input. Client/server describes a particular implementation of distributed computing where several PC's are connected to a server (PC).

Computer networks are often classified by size, distance between locations, topology, media used, etD. There are two general classifications of networks, LAN and WAN. A LAN is a relatively small number of computers, connected by media owned by the company, and generally not exceeding tens of kilometers in size. A WAN often interconnects LANs from various regions of a country or the world, using leased communications lines from third-party vendors. Transmission speeds in LAN topologies is normally measured in megabits per second (mbps), while in WAN environments, speeds are often measured in kilobits per second (kbps). Furthermore, WAN networks are classified as either enterprise, connecting all the LANs of a single organization, or global WAN networks, which interconnect networks of different organizations and spans the earth.

All networks require three elements: individuals with something to share, a pathway to communicate and rules for communication. In modern networks, services are capabilities that computers share. This has given rise to classifying of networks as either peer-to-peer networks, where each computer may request and provide services, or server-centric networks, where roles are more strictly defined; that is, in server-centric networks we generally restrict which computers can request and which computers provide services. We then call computers that provide services servers, and we call computers that request services clients. The second element, a pathway to communicate, includes both wire and wireless transmission media. The third element, rules to communicate, has given rise to protocols.

All networks require three elements: individuals with something to share, a pathway to communicate and rules for communication. In modern networks, services are capabilities that computers share. This has given rise to classifying of networks as either peer-to-peer networks, where each computer may request and provide services, or server-centric networks, where roles are more strictly defined; that is, in server-centric networks we generally restrict which computers can request and which computers provide services. We then call computers that provide services servers, and we call computers that request services clients. The second element, a pathway to communicate, includes both wire and wireless transmission media. The third element, rules to communicate, has given rise to protocols.

All networks require three elements: individuals with something to share, a pathway to communicate and rules for communication. In modern networks, services are capabilities that computers share. Combining these services into a single entity has given rise to the creation of network operating systems (NOS). A NOS generally will run on a designated computer and provide services to clients that communicate requests to the server. The most common services include file, print, message, database and application services. Message services include the storage and delivery of audio, video, text, binary data and graphical content. Message services, in their simplest form, are implemented as electronic mail or e-mail. A more complex but powerful implementation involves integration of e-mail and phone/voice systems. The most sophisticated current implementations are called workflow or workgroup applications. Lotus Notes or GroupWise are examples of high-level message services applications.

Wire transmission media in computer networks conduct either electrical signals or lights. There are three commonly used types of cable media: twisted pair cable, coaxial cable and fiber optic cable. Twisted pair cable consists of insulated 22 to 26 gauge copper wires twisted around each other. The twisting reduces crosstalk and helps cancel signal emissions from the individual wires. Twisted pair cable comes in shielded (STP) and unshielded (UTP), with the shielded providing better immunity from EMI. UTP is generally installed in much the same manner as telephone systems, with the user device connecting to a wall port, the wall port connecting back to one side of a punch-down block, and then connecting to a patch panel. From there small lengths of UTP connect the devices to hubs. Shielded twisted pair does not improve attenuation, but achieves higher throughput rates due to its better immunity from interference. STP is installed used pre-made cable segments, with user devices connected in a star arrangement to a hub (MAU), and the hubs connected to each other in a ring. STP is often associated with IBM token ring, since these were the first systems to use STP on a large scale.

Devices need to be able to connect to the transmission media, and the cable segments need to be connected together to allow for communication to proceed. Devices that are used to connect to media segments include transmission media connectors, network interface boards and modems. Transmission media connectors are attached directly to the media, and allow the computing devices to connect to the media using network interface boards, which in our computer networks is usually a board that is installed into the computer to complete the connection between the computer and the media. Modems are used when computers need to connect to analog networks. Modems convert the computer's digital transmission to analog for transmission, and then convert the incoming analog transmission to digital for the computer. Devices that interconnect separate segments of transmission media include repeaters, hubs, bridges and multiplexors.

Devices need to be able to connect to the transmission media, and the cable segments need to be connected together to allow for communication to proceed. Devices that are used to connect to media segments include transmission media connectors, network interface boards and modems. Devices that interconnect separate segments of transmission media include repeaters, hubs, bridges and multiplexors. Repeaters are used to amplify or regenerate the signal, thereby extending the media's maximum effective length. Repeaters are either amplifiers, which only boost the signal and will also amplify noise, or signal regenerating amplifiers that strips the data from the signal and then re-creates the transmission. Signal regeneration results in cleaner transmissions, but requires more complex equipment and introduces some delay due to the time needed to process the signal. Hubs are used as a central point of connection for multiple segments of media, and include active, passive, multiport repeaters and switches.
Bridges are used to extend the network by connecting separate network segments. Bridges only forward signals along ports if the signal is not intended for any device on the incoming port. They accomplish this by reading the physical device address (MAC address) of the packet, and so can process very quickly providing fast throughput. Because they selectively forward packets, bridges are often used to reduce traffic between segments, thereby reducing network congestion. Multiplexors combine the signals from multiple devices for transmission on a media segment and are used to allow for maximum use of the media's capacity. Routers, brouters and CSU/DSU devices are used to connect multiple independent networks. Because these networks are independent, they are then identified by logical network addresses, forming what are then referred to as subnetworks or subnets. These devices require processing capabilities to identify the network address information inside an incoming packet and determine along which port to forward the packet for eventual delivery to the destination network. These devices generally are slower than their simpler network connectivity cousins, such as bridges, switches, etc.

Routers, brouters and CSU/DSU devices are used to connect multiple independent networks. Because these networks are independent, they are then identified by logical network addresses, forming what are then referred to as subnetworks or subnets. These devices require processing capabilities to identify the network address information inside an incoming packet and determine along which port to forward the packet for eventual delivery to the destination network. These devices generally are slower than their simpler network connectivity cousins, such as bridges, switches, etc. Brouters are similar to routers, but if they do not support the routing protocol, they will bridge the packet.

Devices need to be able to connect to the transmission media, and the cable segments need to be connected together to allow for communication to proceed. Devices that are used to connect to media segments include transmission media connectors, network interface boards and modems. Devices that interconnect separate segments of transmission media include repeaters, hubs, bridges and multiplexors. Repeaters are used to amplify or regenerate the signal, thereby extending the media's maximum effective length. Repeaters are either amplifiers, which only boost the signal and will also amplify noise, or signal regenerating amplifiers that strips the data from the signal and then re-creates the transmission. Signal regeneration results in cleaner transmissions, but requires more complex equipment and introduces some delay due to the time needed to process the signal. Hubs are used as a central point of connection for multiple segments of media, and include active, passive, multiport repeaters and switches. Bridges are used to extend the network by connecting separate network segments. Bridges only forward signals along ports if the signal is not intended for any device on the incoming port. They accomplish this by reading the physical device address (MAC address) of the packet, and so can process very quickly providing fast throughput. Because they selectively forward packets, bridges are often used to reduce traffic between segments, thereby reducing network congestion. Multiplexors combine the signals from multiple devices for transmission on a media segment and are used to allow for maximum use of the media's capacity. Routers, brouters and CSU/DSU devices are used to connect multiple independent networks. Because these networks are independent, they are then identified by logical network addresses, forming what are then referred to as subnetworks or subnets. These devices require processing capabilities to identify the network address information inside an incoming packet and determine along which port to forward the packet for eventual delivery to the destination network. These devices generally are slower than their simpler network connectivity cousins, such as bridges, switches, etc.

The OSI model was developed by the International Standards Organization (ISO) to better describe and understand how networking and communications protocols work. The OSI model divides networking tasks into seven layers: physical, data-link, network, transport, session, presentation and application. Implementations that provide for communications are called protocols. When communications occurs, data is passed down the sending entity's protocol stack layer by layer, with each layer adding information, called headers, to be used at the receiving entity. Layers communicate with each other through predefined addresses called service access points. The sending and receiving entities then interact by sending and receiving messages through headers at peer layers.

The OSI model was developed by the International Standards Organization (ISO) to better describe and understand how networking and communications protocols work. The OSI model divides networking tasks into seven layers: physical, data-link, network, transport, session, presentation and application. Implementations that provide for communications are called protocols. A protocol stack is a hierarchical group of individual protocols that work together. When communications occurs, data is passed down the sending entity's protocol stack layer by layer, with each layer adding information, called headers, to be used at the receiving entity. Layers communicate with each other through predefined addresses called service access points. The sending and receiving entities then interact by sending and receiving messages through headers at peer layers.

The OSI model was developed by the International Standards Organization (ISO) to better describe and understand how networking and communications protocols work. The OSI model divides networking tasks into seven layers: physical, data-link, network, transport, session, presentation and application. Implementations that provide for communications are called protocols. When communications occurs, data is passed down the sending entity's protocol stack layer by layer, with each layer adding information, called headers, to be used at the receiving entity. Layers communicate with each other through predefined addresses called service access points. The sending and receiving entities then interact by sending and receiving messages through headers at peer layers.

The OSI model was developed by the International Standards Organization (ISO) to better describe and understand how networking and communications protocols work. The OSI model divides networking tasks into seven layers: physical, data-link, network, transport, session, presentation and application. Implementations that provide for communications are called protocols. When communications occurs, data is passed down the sending entity's protocol stack layer by layer, with each layer adding information, called headers, to be used at the receiving entity. Layers communicate with each other through predefined addresses called service access points. The sending and receiving entities then interact by sending and receiving messages through headers at peer layers.

Each layer in the OSI model accepts data from its upper layer, appends a control header, and then passes the resulting packet to its lower layer. Each layer data packet has a name specific to that layer. Physical - Bits & packets, Data-link - Frames & packets, Network - Datagrams & packets, Transport - Datagrams, Segments & packets, Session - Packets, Presentation - Packets, and Application - Messages & packets.

Each layer in the OSI model accepts data from its upper layer, appends a control header, and then passes the resulting packet to its lower layer. Each layer data packet has a name specific to that layer. Physical - Bits & packets, Data-link - Frames & packets, Network - Datagrams & packets, Transport - Datagrams, Segments & packets, Session - Packets, Presentation - Packets, and Application - Messages & packets.

Data communication between entities begins as headers are added to the data received from the upper layer and then that header and data are passed to the lower layer as data, and so on until the data is received by the receiving entity. Then the process is reversed, with headers being stripped off as each layer performs according to the instructions contained before passing the data to the upper layer.

Devices that are used to connect to media segments include transmission media connectors, network interface boards and modems, while devices that interconnect separate segments of transmission media include repeaters, hubs, bridges and multiplexors. Routers, brouters and CSU/DSU devices are used to connect multiple independent networks. Because these networks are independent, they are then identified by logical network addresses, forming what are then referred to as subnetworks or subnets. Repeaters, hubs, transmission media adaptors and modems are normally considered physical layer devices, while bridges, switches and network interface cards are normally considered data link layer devices. Routers, brouters and CSU/DSUs are normally considered network layer devices.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. Within connection types there are only two methods: point-to-point and multipoint. Point-to-point connections involve a direct link between two devices. A printer connected to a computer is an example of a point-to-point connection. Multipoint connections involve multiple devices connected, often in star, bus or cellular physical topologies.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. Within connection types there are only two methods: point-to-point and multipoint. Point-to-point connections involve a direct link between two devices. A printer connected to a computer is an example of a point-to-point connection. Multipoint connections involve multiple devices connected, often in star, bus or cellular physical topologies.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. Within connection types there are only two methods: point-to-point and multipoint. Point-to-point connections involve a direct link between two devices. A printer connected to a computer is an example of a point-to-point connection. Multipoint connections involve multiple devices connected, often in star, bus or cellular physical topologies. The second process, physical topology, defines five methods for physically connecting devices to transmission media: bus, ring, star, mesh and cellular.
Bus topologies use a long section of media, called a backbone, to which devices connect with taps, using drops or drop cables. As the signal passes each device connection, there is some signal loss, resulting in limitations on the number of devices that can connect while still maintaining a viable signal. The backbone must be grounded, and both ends must be terminated to remove the signal from the wire and eliminate signal reflection.

Bridges are used to extend the network by connecting separate network segments. Bridges only forward signals along ports if the signal is not intended for any device on the incoming port. They accomplish this by reading the physical device address (MAC address) of the packet, and so can process very quickly providing fast throughput. As packets are received by the bridge, it "learns" the port from which the packet came and the sources MAC address. This allows it to build a simple list of devices on each port so that subsequent communications can be filtered. Because they selectively forward packets, bridges are often used to reduce traffic between segments, thereby reducing network congestion.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. Bandwidth use methods are schemes to maximize the use of a media's available bandwidth. Done through the creation of channels, a portion of the media's bandwidth, bandwidth schemes are either baseband, using the entire bandwidth as a single channel, or broadband, dividing the capacity into multiple channels. Bandwidth use schemes employ multiplexing, allowing multiple devices to communicate simultaneously over a single media segment. Multiplexing techniques include frequency-division multiplexing (FDM), time-division (TDM) and statistical time-division multiplexing (StatTDM). (FDM) uses separate frequencies to provide multiple channels on a transmission medium. Cable TV is a good example of an FDM technology. TDM divides a single channel into time slots and provides availability to the pre-set time slots. FDM can be used on broadband and baseband systems, but suffers from wasted bandwidth if slots are not in use. StatTDM is an enhancement of TDM that allows dynamic allocation of time slots on a first-come first-served basis.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. Within connection types there are only two methods: point-to-point and multipoint. Point-to-point connections involve a direct link between two devices. A printer connected to a computer is an example of a point-to-point connection. Multipoint connections involve multiple devices connected, often in star, bus or cellular physical topologies. The second process, physical topology, defines five methods for physically connecting devices to transmission media: bus, ring, star, mesh and cellular. Bus topologies use a long section of media, called a backbone, to which devices connect with taps, using drops or drop cables. As the signal passes each device connection, there is some signal loss, resulting in limitations on the number of devices that can connect while still maintaining a viable signal. The backbone must be grounded, and both ends must be terminated to remove the signal from the wire and eliminate signal reflection.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. The second process, physical topology, defines five methods for physically connecting devices to transmission media: bus, ring, star, mesh and cellular. Ring topologies involve a closed loop of point-to-point connections. Devices either connect directly to the ring or indirectly using an interface card and drop cable. IBM token ring uses hubs arranged in a physical ring, while user devices connect to the hubs in a star topology.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. The second process, physical topology, defines five methods for physically connecting devices to transmission media: bus, ring, star, mesh and cellular. Cellular topologies are wireless systems that allow devices to move around within geographic areas, called cells, as they maintain communications with hubs. The hubs are then interconnected providing for routing of communications from cell to cell.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. The third and fourth processes, digital and analog signaling, describe methods for using electrical energy to represent data. This process is called signaling or encoding or modulation. In digital signaling, two forms of modulation are described: current state and state transition. Current state methods measure the presence or absence of a characteristic of a signal. Using light sources in fiber optic transmission is a good example of current state signaling. State transition methods measure the transition from one state to another, rather than the absolute presence or absence.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. The third and fourth processes, digital and analog signaling, describe methods for using electrical energy to represent data. This process is called signaling or encoding or modulation. In digital signaling, two forms of modulation are described: current state and state transition. Current state methods measure the presence or absence of a characteristic of a signal. Using light sources in fiber optic transmission is a good example of current state signaling. State transition methods measure the transition from one state to another, rather than the absolute presence or absence. In analog signaling, three methods are commonly employed: amplitude-shift keying (ASK), frequency-shift keying (FSK), and phase-shift keying (PSK). The first two are considered current state methods, while PSK is a state transition method.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. Bit synchronization is a process by which the receiving device and the sending device know when to measure the specific characteristic of the transmission that is being modulated to represent data. Knowing when to measure allows both devices to communicate with either current state or state transition schemes. Bit synchronization is either asynchronous or synchronous. Asynchronous methods use start and stop bits, and devices must provide their own internal clocks for measuring time. Synchronous methods use strategies to provide clocking for devices. Synchronous methods include guaranteed state change, which embeds a clock signal in the data transmission, separate clock signal, which provides a separate channel for clocking, and oversampling, where the receiver samples the data signal at a much faster rate than it was sent.

The physical layer of the OSI model lists seven processes, each with its own methods, to enable communication at this level. The seven processes are connection types, physical topology, digital signaling, analog signaling, bit synchronization, bandwidth use and multiplexing. Bit synchronization is a process by which the receiving device and the sending device know when to measure the specific characteristic of the transmission that is being modulated to represent data. Knowing when to measure allows both devices to communicate with either current state or state transition schemes. Bit synchronization is either asynchronous or synchronous. Asynchronous methods use start and stop bits, and devices must provide their own internal clocks for measuring time. Synchronous methods use strategies to provide clocking for devices. Synchronous methods include guaranteed state change, which embeds a clock signal in the data transmission, separate clock signal, which provides a separate channel for clocking, and oversampling, where the receiver samples the data signal at a much faster rate than it was sent.

With separate clock signaling, one channel is used for data and the other is used to provide clocking, thus doubling the needed bandwidth for communications. Bit synchronization is a process by which the receiving device and the sending device know when to measure the specific characteristic of the transmission that is being modulated to represent data. Knowing when to measure allows both devices to communicate with either current state or state transition schemes. Bit synchronization is either asynchronous or synchronous. Asynchronous methods use start and stop bits, and devices must provide their own internal clocks for measuring time. Synchronous methods use strategies to provide clocking for devices. Synchronous methods include guaranteed state change, which embeds a clock signal in the data transmission, separate clock signal, which provides a separate channel for clocking, and oversampling, where the receiver samples the data signal at a much faster rate than it was sent.

Bandwidth use schemes employ multiplexing, allowing multiple devices to communicate simultaneously over a single media segment. Multiplexing techniques include frequency-division multiplexing (FDM), time-division (TDM) and statistical time-division multiplexing (StatTDM). (FDM) uses separate frequencies to provide multiple channels on a transmission medium. Cable TV is a good example of an FDM technology. TDM divides a single channel into time slots and provides availability to the pre-set time slots. FDM can be used on broadband and baseband systems, but suffers from wasted bandwidth if slots are not in use. StatTDM is an enhancement of TDM that allows dynamic allocation of time slots on a first-come first-served basis.

The second layer of the OSI model is the data link layer. Protocols at the data link layer are responsible for organizing the physical layer bits into frames for transmission, error detection, data flow control and physical device addressing. Devices normally associated with this layer include bridges, switches and NIC cards. There are two sub-layers. The MAC sublayer describes processes for addressing, media access and logical topology. Addressing is physical device addressing, as seen commonly in NIC cards as the MAC address that is burned into the card by the manufacturer. Media access methods describe how devices share the transmission medium and include contention, token-passing and polling. Logical topologies describe the actual path a signal takes, which can be different than the physical topology.

The second layer of the OSI model is the data link layer. Protocols at the data link layer are responsible for organizing the physical layer bits into frames for transmission, error detection, data flow control and physical device addressing. Devices normally associated with this layer include bridges, switches and NIC cards. There are two sub-layers. The MAC sublayer describes processes for addressing, media access and logical topology. Addressing is physical device addressing, as seen commonly in NIC cards as the MAC address that is burned into the card by the manufacturer. Media access methods describe how devices share the transmission medium and include contention, token-passing and polling. Logical topologies describe the actual path a signal takes, which can be different than the physical topology.

The second layer of the OSI model is the data link layer. Protocols at the data link layer are responsible for organizing the physical layer bits into frames for transmission, error detection, data flow control and physical device addressing. Devices normally associated with this layer include bridges, switches and NIC cards. There are two sub-layers. The MAC sublayer describes processes for addressing, media access and logical topology. Addressing is physical device addressing, as seen commonly in NIC cards as the MAC address that is burned into the card by the manufacturer. Media access methods describe how devices share the transmission medium and include contention, token-passing and polling. Logical topologies describe the actual path a signal takes, which can be different than the physical topology. Logical topology can be both bus, where the signal is "broadcast" throughout the network, regardless of where the destination device is, and ring, where the signal always travels in a specified path in a single direction.

The second layer of the OSI model is the data link layer. Protocols at the data link layer are responsible for organizing the physical layer bits into frames for transmission, error detection, data flow control and physical device addressing. Devices normally associated with this layer include bridges, switches and NIC cards. There are two sub-layers. The MAC sublayer describes processes for addressing, media access and logical topology. Addressing is physical device addressing, as seen commonly in NIC cards as the MAC address that is burned into the card by the manufacturer. Media access methods describe how devices share the transmission medium and include contention, token-passing and polling. Logical topologies describe the actual path a signal takes, which can be different than the physical topology.
Of the three media access methods, contention is the one PC networks use most commonly. Contention is a first-come first-served method, which allows for simple connectivity devices and fast access, but as the number of devices increases, then the probability of two devices communicating simultaneously increases. The result of this simultaneous transmission is a collision, where the data packets are damaged and must be retransmitted. Newer implementations use carrier sense, multiple access (CSNMA) schemes, where transmitting devices listen to the media to determine if they can transmit. The most common CSMA protocols use collision detection (CSMA/CD), where a third device detecting a collision will send a "jamming" signal out, and all devices will then wait a random amount of time before again trying to transmit. In token passing, a small packet called a token is passed from device to device. Access to the media is only available when the device has the token; this method eliminates the potential for collisions, and so performs better than contention as the device population increases. Polling systems have one device that polls the other devices and gives them access one-at-a-time. Polling systems are used most often in time-sensitive networking environments.

The second layer of the OSI model is the data link layer. Protocols at the data link layer are responsible for organizing the physical layer bits into frames for transmission, error detection, data flow control and physical device addressing. Devices normally associated with this layer include bridges, switches and NIC cards. There are two sub-layers. The MAC sublayer describes processes for addressing, media access and logical topology. Addressing is physical device addressing, as seen commonly in NIC cards as the MAC address that is burned into the card by the manufacturer. Media access methods describe how devices share the transmission medium and include contention, token-passing and polling. Logical topologies describe the actual path a signal takes, which can be different than the physical topology.
Of the three media access methods, contention is the one PC networks use most commonly. Contention is a first-come first-served method, which allows for simple connectivity devices and fast access, but as the number of devices increases, then the probability of two devices communicating simultaneously increases. The result of this simultaneous transmission is a collision, where the data packets are damaged and must be retransmitted. Newer implementations use carrier sense, multiple access (CSNMA) schemes, where transmitting devices listen to the media to determine if they can transmit. The most common CSMA protocols use collision detection (CSMA/CD), where a third device detecting a collision will send a "jamming" signal out, and all devices will then wait a random amount of time before again trying to transmit. In token passing, a small packet called a token is passed from device to device. Access to the media is only available when the device has the token; this method eliminates the potential for collisions, and so performs better than contention as the device population increases. Polling systems have one device that polls the other devices and gives them access one-at-a-time. Polling systems are used most often in time-sensitive networking environments.

The second layer of the OSI model is the data link layer. Protocols at the data link layer are responsible for organizing the physical layer bits into frames for transmission, error detection, data flow control and physical device addressing. Devices normally associated with this layer include bridges, switches and NIC cards. There are two sub-layers. The MAC sublayer describes processes for addressing, media access and logical topology. Addressing is physical device addressing, as seen commonly in NIC cards as the MAC address that is burned into the card by the manufacturer. Media access methods describe how devices share the transmission medium and include contention, token-passing and polling. Logical topologies describe the actual path a signal takes, which can be different than the physical topology.
Of the three media access methods, contention is the one PC networks use most commonly. Contention is a first-come first-served method, which allows for simple connectivity devices and fast access, but as the number of devices increases, then the probability of two devices communicating simultaneously increases. The result of this simultaneous transmission is a collision, where the data packets are damaged and must be retransmitted. Newer implementations use carrier sense, multiple access (CSNMA) schemes, where transmitting devices listen to the media to determine if they can transmit. The most common CSMA protocols use collision detection (CSMA/CD), where a third device detecting a collision will send a "jamming" signal out, and all devices will then wait a random amount of time before again trying to transmit. In token passing, a small packet called a token is passed from device to device. Access to the media is only available when the device has the token; this method eliminates the potential for collisions, and so performs better than contention as the device population increases. Token passing is often used in time-sensitive network environments. Polling systems have one device that polls the other devices and gives them access one-at-a-time. Polling systems are also used in time-sensitive networking environments.

The second layer of the OSI model is the data link layer. Devices normally associated with this layer include bridges, switches and NIC cards. Bridges are used to extend the network by connecting separate network segments. Bridges only forward signals along ports if the signal is not intended for any device on the incoming port. They accomplish this by reading the physical device address (MAC address) of the packet, and so can process very quickly providing fast throughput. As packets are received by the bridge, it "learns" the port from which the packet came and the sources MAC address. This allows it to build a simple list of devices on each port so that subsequent communications can be filtered. Because they selectively forward packets, bridges are often used to reduce traffic between segments, thereby reducing network congestion.

The second layer of the OSI model is the data link layer. Protocols at the data link layer are responsible for organizing the physical layer bits into frames for transmission, error detection, data flow control and physical device addressing. Devices normally associated with this layer include bridges, switches and NIC cards. There are two sub layers of the data link layer, the MAC and LLC. The LLC sublayer is responsible for transmission synchronization and connection services. Transmission synchronization methods include asynchronous, synchronous and isochronous. Asynchronous methods require that each device maintain its own timing clock. Transmissions are sent with start and stop bits, and a parity bit is generally added for error checking. Synchronous methods require the devices to provide clocking help, either in the form of special strings (SYNC bits) or the use of a separate channel for timing signals, and use cyclic redundancy checking (CRC) for error detection. Synchronous methods allow for larger amounts of data transfer with less start and stop bits, thereby reducing overhead and increasing throughput rates. Isochronous methods use a transmission clock to create time slots for transmission and provide timing signals for transmission.

The second layer of the OSI model is the data link layer. Protocols at the data link layer are responsible for organizing the physical layer bits into frames for transmission, error detection, data flow control and physical device addressing. Devices normally associated with this layer include bridges, switches and NIC cards. There are two sub layers of the data link layer, the MAC and LLC. The LLC sublayer is responsible for transmission synchronization and connection services. Transmission synchronization methods include asynchronous, synchronous and isochronous. Asynchronous methods require that each device maintain its own timing clock. Transmissions are sent with start and stop bits, and a parity bit is generally added for error checking. Synchronous methods require the devices to provide clocking help, either in the form of special strings (SYNC bits) or the use of a separate channel for timing signals, and use cyclic redundancy checking (CRC) for error detection. Synchronous methods allow for larger amounts of data transfer with less start and stop bits, thereby reducing overhead and increasing throughput rates. Isochronous methods use a transmission clock to create time slots for transmission and provide timing signals for transmission.

The second layer of the OSI model is the data link layer. Protocols at the data link layer are responsible for organizing the physical layer bits into frames for transmission, error detection, data flow control and physical device addressing. Devices normally associated with this layer include bridges, switches and NIC cards. There are two sub layers of the data link layer, the MAC and LLC. The LLC sublayer is responsible for transmission synchronization and connection services. Transmission synchronization methods include asynchronous, synchronous and isochronous. Connection services include unacknowledged connectionless, connection-oriented and acknowledged connectionless services. These services are defined at several layers of the OSI model. Using special packets called acknowledgements, connection services allow for flow control and error control at the data link layer. Flow control involves either guaranteed rate or window methods. Guaranteed rate establishes an acceptable transmission rate between sender ands receiver before the transmissions begin, and lasting until the transmissions cease. Window flow control establishes a window or buffer of frames that can be transmitted. Window flow control can be either static, where the number is established at connection and does not change, or dynamic. In dynamic flow control, the sending entity will increase the rate of transmission until the receiving entity sends a "choke" packet, telling the sender to slow the rate down. Error control at this sublayer can include buffer overflow errors, packet or frame size errors, signal noise problems, mismatching checksums and negative acknowledgements.

Of the three media access methods, contention is the one PC networks use most commonly. Contention is a first-come first-served method, which allows for simple connectivity devices and fast access, but as the number of devices increases, then the probability of two devices communicating simultaneously increases. The result of this simultaneous transmission is a collision, where the data packets are damaged and must be retransmitted. Newer implementations use carrier sense, multiple access (CSNMA) schemes, where transmitting devices listen to the media to determine if they can transmit. The most common CSMA protocols use collision detection (CSMA/CD), where a third device detecting a collision will send a "jamming" signal out, and all devices will then wait a random amount of time before again trying to transmit.

Media access methods describe how devices share the transmission medium and include contention, token-passing and polling. Logical topologies describe the actual path a signal takes, which can be different than the physical topology. Of the three media access methods, contention is the one PC networks use most commonly. Contention is a first-come first-served method, which allows for simple connectivity devices and fast access, but as the number of devices increases, then the probability of two devices communicating simultaneously increases. The result of this simultaneous transmission is a collision, where the data packets are damaged and must be retransmitted. Newer implementations use carrier sense, multiple access (CSNMA) schemes, where transmitting devices listen to the media to determine if they can transmit. The most common CSMA protocols use collision detection (CSMA/CD), where a third device detecting a collision will send a "jamming" signal out, and all devices will then wait a random amount of time before again trying to transmit. In token passing, a small packet called a token is passed from device to device. Access to the media is only available when the device has the token. This method eliminates the potential for collisions, and so performs better than contention as the device population increases. Token passing is often used in time-sensitive network environments, because access can be prioritized. Polling systems have one device that polls the other devices and gives them access one-at-a-time. The polling device can be instructed to give priority to certain devices. Polling systems are also used in time-sensitive networking environments.

The OSI model was developed by the International Standards Organization (ISO) to better describe and understand how networking and communications protocols work. The OSI model divides networking tasks into seven layers: physical, data-link, network, transport, session, presentation and application. Implementations that provide for communications are called protocols. The network layer defines six processes, addressing, switching, route discovery, route selection, connection services and gateway services. Addressing at the network layer includes logical network addresses and service addresses. Logical network addresses are used to distinguish between two different networks in an internetwork. Often referred to as subnetworks or subnets, these addresses are assigned by the administrator, based upon the addressing scheme of the particular protocol implementation in use. Service addresses are also called ports or sockets, and are used to identify a specific upper-layer protocol or service.

The OSI model was developed by the International Standards Organization (ISO) to better describe and understand how networking and communications protocols work. The OSI model divides networking tasks into seven layers: physical, data-link, network, transport, session, presentation and application. Implementations that provide for communications are called protocols. The network layer defines six processes, addressing, switching, route discovery, route selection, connection services and gateway services. Addressing at the network layer includes logical network addresses and service addresses. Logical network addresses are used to distinguish between two different networks in an internetwork. Often referred to as subnetworks or subnets, these addresses are assigned by the administrator, based upon the addressing scheme of the particular protocol implementation in use. Service addresses are also called ports or sockets, and are used to identify a specific upper-layer protocol or service.

The OSI model was developed by the International Standards Organization (ISO) to better describe and understand how networking and communications protocols work. The OSI model divides networking tasks into seven layers: physical, data-link, network, transport, session, presentation and application. Implementations that provide for communications are called protocols. The network layer defines six processes, addressing, switching, route discovery, route selection, connection services and gateway services. Addressing at the network layer includes logical network addresses and service addresses. Logical network addresses are used to distinguish between two different networks in an internetwork. Often referred to as subnetworks or subnets, these addresses are assigned by the administrator, based upon the addressing scheme of the particular protocol implementation in use. Service addresses are also called ports or sockets, and are used to identify a specific upper-layer protocol or service.

The OSI model was developed by the International Standards Organization (ISO) to better describe and understand how networking and communications protocols work. The OSI model divides networking tasks into seven layers: physical, data-link, network, transport, session, presentation and application. Implementations that provide for communications are called protocols. The network layer defines six processes, addressing, switching, route discovery, route selection, connection services and gateway services. Switching techniques include circuit switching, message switching and packet switching. Circuit switching establishes a circuit for the duration of the conversation, providing a dedicated path for communications. Telephone calls are an example of this technique. Message switching breaks the conversation into smaller pieces, called messages, and each message must be sent through the internetwork to the destination device. In message switching, intermediate devices must store the message for a brief period of time, depending upon the network conditions, and then forward it on. These are often referred to as store-and-forward networks, and devices need storage devices to function. Packet switching breaks messages into small packets that can be stored in RAM rather than on hard drives. Packet switching methods are virtual circuit or datagram packet switching. Virtual circuits are logical connections established between sender and receiver at the beginning of the conversation setting parameters such as path to be taken, maximum message size, etc. Virtual circuits appear to the communicating entities as point-to-point links. Datagram packet switching does not use logical connections, but rather lets the devices in the internetwork determine the path dynamically. The Internet uses datagram packet switching.

The network layer defines six processes, addressing, switching, route discovery, route selection, connection services and gateway services. Switching techniques include circuit switching, message switching and packet switching. Circuit switching establishes a circuit for the duration of the conversation, providing a dedicated path for communications. Telephone calls are an example of this technique. Message switching breaks the conversation into smaller pieces, called messages, and each message must be sent through the internetwork to the destination device. In message switching, intermediate devices must store the message for a brief period of time, depending upon the network conditions, and then forward it on. These are often referred to as store-and-forward networks, and devices need storage devices to function. Packet switching breaks messages into small packets that can be stored in RAM rather than on hard drives. Packet switching methods are virtual circuit or datagram packet switching. Virtual circuits are logical connections established between sender and receiver at the beginning of the conversation setting parameters such as path to be taken, maximum message size, etc. Virtual circuits appear to the communicating entities as point-to-point links. Datagram packet switching does not use logical connections, but rather lets the devices in the internetwork determine the path dynamically. The Internet uses datagram packet switching.

The network layer defines six processes, addressing, switching, route discovery, route selection, connection services and gateway services. Switching techniques include circuit switching, message switching and packet switching. Circuit switching establishes a circuit for the duration of the conversation, providing a dedicated path for communications. Telephone calls are an example of this technique. Message switching breaks the conversation into smaller pieces, called messages, and each message must be sent through the internetwork to the destination device. In message switching, intermediate devices must store the message for a brief period of time, depending upon the network conditions, and then forward it on. These are often referred to as store-and-forward networks, and devices need storage devices to function. Packet switching breaks messages into small packets that can be stored in RAM rather than on hard drives. Packet switching methods are virtual circuit or datagram packet switching. Virtual circuits are logical connections established between sender and receiver at the beginning of the conversation setting parameters such as path to be taken, maximum message size, etc. Virtual circuits appear to the communicating entities as point-to-point links. Datagram packet switching does not use logical connections, but rather lets the devices in the internetwork determine the path dynamically. The Internet uses datagram packet switching.

The OSI model was developed by the International Standards Organization (ISO) to better describe and understand how networking and communications protocols work. The OSI model divides networking tasks into seven layers: physical, data-link, network, transport, session, presentation and application. Implementations that provide for communications are called protocols. The network layer defines six processes, addressing, switching, route discovery, route selection, connection services and gateway services. Route discovery methods identify the means by which we identify routes and maintain route tables. Route tables list the known networks, including the next device in the path and costs using hop (the number of routers to the destination network) and tick count (a time required to reach the destination network). There are two types of route discovery methods, distance vector and link-state. Distance vector routers maintain their route tables by periodically broadcasting their entire table to their neighbors, and receiving their neighbor's tables. When the router receives the broadcast from its neighbors, it then must recalculate its route table. While distance vector methods are simple and require little management, they create lots of traffic and have problems with convergence as the number of routers increases. Link state methods only call for notification if a change occurs, resulting in much less traffic.

Route discovery methods identify the means by which we identify routes and maintain route tables. Route tables list the known networks, including the next device in the path and costs using hop (the number of routers to the destination network) and tick count (a time required to reach the destination network). There are two types of route discovery methods, distance vector and link-state. Distance vector routers maintain their route tables by periodically broadcasting their entire table to their neighbors, and receiving their neighbor's tables. When the router receives the broadcast from its neighbors, it then must recalculate its route table. While distance vector methods are simple and require little management, they create lots of traffic and have problems with convergence as the number of routers increases. Link state methods only call for notification if a change occurs, resulting in much less traffic.

Route discovery methods identify the means by which we identify routes and maintain route tables. Route tables list the known networks, including the next device in the path and costs using hop (the number of routers to the destination network) and tick count (a time required to reach the destination network). There are two types of route discovery methods, distance vector and link-state. Distance vector routers maintain their route tables by periodically broadcasting their entire table to their neighbors, and receiving their neighbor's tables. When the router receives the broadcast from its neighbors, it then must recalculate its route table. While distance vector methods are simple and require little management, they create lots of traffic and have problems with convergence as the number of routers increases. Link state methods only call for notification if a change occurs, resulting in much less traffic.

Network layer flow control implementations are designed to change with the internetworks capabilities. Because of the size and complexity of internetworks, multiple routes often exist to the destination network, while traffic and device status changes moment by moment. Devices use acknowledgements and negative acknowledgements at this layer to detect problems and communicate failures. PING, from TCP/IP, is an example of a process used to check internetwork conditions for congestion. Network layer flow control can be static or dynamic, similar to the data link layer. Flow control also involves intelligent path selection (routing), and so flow control at the network layer is also called congestion control. Error control at this layer deals with dropped or lost packets, duplicate packets and bad data (as identified through checksums).

The OSI transport layer is designed to hide the intricacies of the network from the upper layer processes, handle segmenting of upper-layer messages into smaller segments that the lower-layer processes can handle, while combining small lower-layer segments into messages for the upper layers, and to help compensate for the lack of reliable services in the lower layers. There are four processes described at this layer: address-name resolution, addressing, segment development and connection services.

The OSI transport layer is designed to hide the intricacies of the network from the upper layer processes, handle segmenting of upper-layer messages into smaller segments that the lower-layer processes can handle, while combining small lower-layer segments into messages for the upper layers, and to help compensate for the lack of reliable services in the lower layers. There are four processes described at this layer: address-name resolution, addressing, segment development and connection services.

The OSI transport layer is designed to hide the intricacies of the network from the upper layer processes, handle segmenting of upper-layer messages into smaller segments that the lower-layer processes can handle, while combining small lower-layer segments into messages for the upper layers, and to help compensate for the lack of reliable services in the lower layers. There are four processes described at this layer: address-name resolution, addressing, segment development and connection services. Address-name resolution consists of either service requestor initiated or service provider initiated methods. A DNS request from a workstation would be an example of a service requestor initiated, while SAP from IPX (Service Advertising Protocol) would be similar to service provider initiated address name resolution methods.

The OSI transport layer is designed to hide the intricacies of the network from the upper layer processes, handle segmenting of upper-layer messages into smaller segments that the lower-layer processes can handle, while combining small lower-layer segments into messages for the upper layers, and to help compensate for the lack of reliable services in the lower layers. There are four processes described at this layer: address-name resolution, addressing, segment development and connection services. Address-name resolution consists of either service requestor initiated or service provider initiated methods. A DNS request from a workstation would be an example of a service requestor initiated, while SAP from IPX (Service Advertising Protocol) would be similar to service provider initiated address name resolution methods.

The OSI transport layer is designed to hide the intricacies of the network from the upper layer processes, handle segmenting of upper-layer messages into smaller segments that the lower-layer processes can handle, while combining small lower-layer segments into messages for the upper layers, and to help compensate for the lack of reliable services in the lower layers. There are four processes described at this layer: address-name resolution, addressing, segment development and connection services. The second process, addressing, consists of either connection identifiers or transaction identifiers. Connection identifiers (connection ID, port socket) identify specific conversations being maintained at a given service address. Transaction identifiers (Transaction ID) refer to a subset of a conversation, usually thought of as a request and a response.