To communicate over telephone lines, computers need a device that will convert the digital signals used by computers into the analog signals used by the telephone system. This device is called a modem, which is a word made up of the two words modulator and demodulator. The modem at the sending end modulates the computer's digital signals into an analog wave, and the modem on the receiving computer demodulates the analog signal back into a digital signal that can be understood by the receiving computer.

Modems are known as data communications equipment (DCE) and use an RS-232 communications interface (serial port) and an RJ-11 telephone plug (four wires). Modems are available in both internal and external models. Internal modems are plugged into a computer's internal expansion bus. External modems connect to the RS-232 serial port of a computer.

The Hayes Microcomputer Products, Inc. invented a product known as the Hayes SmartModem in the early 1980's. It could automatically dial a number through a phone that was hung up. The Hayes modem became the standard and most other modems emulated the Hayes modem.

The current modem standards are set by the International Telecommunications Union (ITU). (See chart on page 538).

Baud refers to the speed of the oscillation of the sound wave on which a bit of data is carried over the phone lines. The term bits per second (BPS) refers to the actual number of bits that can be encoded and transmitted per second. The BPS can be greater than the baud rate because of data compression. The baud rate never exceeds 2400 baud, but the bit rate, including data compression, can be as high as 115,200 BPS (with the latest V.90 standard).

Modems may be roughly divided into two categories, asynchronous, and synchronous. The type of modem that a network uses will depend on the network environment and what the network needs to do.

Each character is translated into a stream of bits. Each stream of bits is separated from other streams by a start bit and a stop bit. The computers must agree on a start/stop bit sequence. Communication is not synchronized. The receiving computer uses the start and stop information to schedule the timing functions so it is ready to receive the next data stream. Twenty-five percent of the data traffic in asynchronous communications consists of data traffic control and coordination.

Microcom developed Microcom Networking Protocol (MNP) because the original V.32 standard di not include error control. Other companies have adopted the MNP 2--4 standards. The CCITT developed the V.42 standard, which employs two hardware-based error control protocols, Link Access Procedure for Modems (LAPM) and MNP 4.

Communications performance depends on two elements:

1. Signaling of channel speed -- this parameter describes how fast the bits are encoded onto the communications channel.
2. Throughput - a measure of the amount of useful information going across the channel.

Using data compression can increase the throughput. MNP Class 5 data compression can double the transmission speed and the V.42bis can quadruple the transmission speed.

Synchronous transmission uses a timing scheme to coordinate the sending and receiving of data. This scheme separates groups of bits into frames, and uses special characters to signal the start and end of a frame. The bits are sent and received in a timed manner, so the transmission speed can be higher and start and stop bits are not required, so synchronous transmission is more efficient than asynchronous transmission.

Synchronous protocols perform a number of tasks that asynchronous protocols do not. Synchronous protocols format data into blocks, add control information, and check the information to provide error control. The primary protocols in synchronous communications are:

• Synchronous data link control
• High-level data link control
• Binary synchronous communications protocol (bisynch)

Synchronous communications are used in almost all digital and network communications.

It is generally considered difficult and expensive to move data quickly over long distances. The three factors that must be considered to when choosing how to implement modem communications are throughput, distance, and cost.

There are two types of telephone lines available for modem communications, public dial-up lines (dial-up lines) and leased lines.

Public dial-up lines are the common telephone lines. These are not the best choice for data transfer as they are slow and not totally reliable. However, they are inexpensive and may be used anywhere that there are dial-up lines (most places in the world). These lines are also know as switched lines because each connection runs through the carrier's switching network; you are unlikely to use exactly the same route each time you dial.

Leased lines are full-time dedicated lines that do not use the switching network to complete the connection. The quality of the line is higher (guaranteed by the carrier) than switched lines and can support higher data transfer rates.

Most networks allow authorized users to connect to the network via telephone lines. Windows NT uses Remote Access Service (RAS), which is a service built in to Windows NT. RAS permits up to 256 clients to dial in. Once the user has made a connection, the user can access all of the network resources as if they were sitting at (an extremely slow) computer at the network site.

Point-to-Point Tunneling Protocol (PPTP) (Page 547) (Read from text)

Repeaters operate at the Physical layer of the OSI model. Repeaters regenerate weak signals and send them out on other segments. To pass the data onto the next segment, the packets and the LLC protocols must be the same on both segments. Repeaters cannot connect cables with different access methods. Repeaters can move packets from one physical media to another, such as 10Base2 to fiber optic, if the repeater supports both cable types.

Some multiport repeaters act as multiport hubs and connect different media. The same segment limits apply, but the limits now refer to each segment extending from the hub rather than the entire network.

Repeaters improve performance by dividing the network into segments. Repeaters are relatively inexpensive. Consider using a repeater to link segments of a network when there is not a lot of traffic on either segment and when cost is a major consideration.

Repeaters pass everything along to the next segment. There is no filtering of packets; repeaters will pass along a broadcast storm (or anything else) to other segments.

Bridges can function to join segments or workgroup LANs. However, a bridge can also divide a network to isolate traffic or problems. Bridges can be used to: (list on Page 556)

Bridges work at the media access control sub-layer of the data-link layer of the OSI model. Because they work at this layer, bridges do not distinguish between protocols, they simply pass all protocols along the network; it is up to the individual computers to determine which protocols that they can recognize. Bridges are sometimes called media access control layer bridges.

A bridge works on the principle that each network node has its own address. A bridge forwards packets based on the address of the destination node. Bridges "learn" where to forward data.

A bridge builds a routing table based on the source address of packets. As nodes transmit packets, the source address is copied to the routing table. Using this information, the bridge learns which computers are on which segment of the network.

Bridges are often used in large networks that have geographically dispersed segments joined by telephone lines. In a situation where two separate LANs are located great distances apart, they may be connected by using two bridges joined together by telephone lines, usually using a data grade line and synchronous modems.

Bridges work at a higher OSI level than repeaters; bridges have more intelligence than repeaters and can take more data features into account. Bridges regenerate data at the packet level, which means that they can send packets over long distances using a variety of long distance media. Bridges provide better performance than repeaters because there will be fewer computers competing for network resources on two separate segments. There will be fewer packets and fewer collisions. The bridge will pass packets between the two segments, where appropriate.

Bridges may be constructed by using two network cards in a server if the NOS supports this.

In a mixed protocol and media access network, bridges may not be adequate to ensure fast communications among all of the segments. Routers know the address of each segment, and can determine the best path for sending data and filtering broadcast traffic to the local segment.

Routers are able to switch and route packets across multiple networks because they work at the Network layer of the OSI model. They do this by exchanging protocol-specific information between separate networks. (Read third paragraph on page 564)

A router uses a table to determine the destination address for incoming data, including:

• All known network addresses
• How to connect to other networks
• The possible paths between those routers
• The costs of sending data over those paths

The router selects the best route for the data based on costs and available paths.

The term "routing table" has a different meaning for routers and bridges. A router's routing table contains network numbers, while a bridge's routing table contains media access control sub-layer addresses for each node.

Routers require specific network numbers, which allow them to talk to other routers and to local NIC addresses. They cannot talk to remote computers.

As packets are passed from one router to another, Data Link layer source and destination addresses are stripped off and then recreated, which enables a router to route a packet from a TCP/IP Ethernet network to a TCP/IP Token Ring network.

Routers will pass information only if the network address is known. This reduces the amount of network traffic between the networks and allows routers to use these links more efficiently than bridges. Because routers do not pass or even handle every packet, they act as a safety barrier between network segments, which can greatly reduce the amount of traffic on the network and the wait time experienced by users.

Routable Protocols (Page 566) (discuss this)

Unlike bridges, routers support multiple paths between LAN segments and can choose among redundant paths. Routers can link segments using different data packing and media access schemes. This means that there will often be several paths available for the router to use. If one router will not function the data can still be sent over an alternate path.

Routers can listen to the network, identify which paths are the busiest, and send data over the least busy path. A router determines which path over which to send the data by determining the number of hops between internetwork segments. Routers build routing tables and use them in these routing algorithms:

OPSF (open shortest path first) is a link-state algorithm. Link-state algorithms control the routing process and allow routers to respond quickly to changes in the network. Link-state routing uses the Dijkstra algorithm to calculate routes based on the number of hops, the line speed, traffic, and cost. TCP/IP supports OSPF.

RIP (routing information protocol) uses distance-vector algorithms to determine routes. TCP/IP and IPX support RIP.

NLPX (NetWare link services protocol is a link-state algorithm used with IPX.

Static routers allow the administrator to manually set up and configure the routing table to specify each route.

Dynamic routers perform an automatic discovery of routes and require little setup and configuration. They examine the information from other routers and make packet-by-packet decisions as to how to send the data across the network.

Forwarding the packet is the key to understanding the differences between bridges and routers. Bridges forward broadcast data to every port of the bridge except the port from which the data was sent. In large networks, this creates enough broadcast traffic to slow the network down despite filtering for network addresses.

The router, which works at the Network layer, takes more information into account, determining not only what to forward, but where to forward it. The router recognizes not only the address, but a type of protocol as well. The router can also identify addresses of other routers and determine which packets to forward to which routers.

A brouter combines the best qualities of bridges and routers. They can act like a router for one protocol and a bridge for all others. Brouters can route selected routable protocols, bridge non-routable protocols, and deliver more cost-effective and more manageable internetworking than separate bridges and routers.

Gateways make communication possible between two different architectures and a environments, repackaging information to meet the requirements of the destination system. They can change the format of a message so that it will conform to the application program on the receiving end of the transfer. A gateway links two systems that do not use the same:

• Communications protocols
• Data formatting structures
• Languages
• Architecture

Gateways are task-specific, which means that they are dedicated to a particular type of transfer. A gateway takes data from one environment, strips off its old protocol stack, and repackages it in the protocol stack from the destination network. To process the data, the gateway:

• Decapsulates the incoming data through the network’s complete protocol stack.
• Encapsulates the outgoing data in the complete protocol stack of the other network to allow transmission.

Some gateways use all 7 layers of the OSI model, but gateways typically perform protocol conversion at the Application layer. However, the level of functionality varies widely between types of gateways.

Gateways are often used to translate between a personal computer network and a mainframe computer. A host gateway connects LAN computers with a mainframe or minicomputer system that doesn’t recognize intelligent computers connected to a LAN.

In a LAN environment, one computer is usually designated as the gateway computer. Applications programs access the mainframe by communicating with the mainframe through the gateway computer. Users can access resources on the mainframe computer just as if those resources were on their own desktop computers.

Gateways are typically dedicated servers on the network because they can use a significant portion of a server’s available bandwidth. They perform resource-intensive tasks such as protocol conversion. Use gateways when different environments need to communicate.

The world-wide public switched telephone network (PSTN) is available for computer communications and can be thought of as one large WAN link. However, these communications require modems, which can make them perform slowly. The network connections do not have consistent quality because the PSTN is a switched network and the routing of the switches varies from connection to connection.

Dedicated analog lines, known as leased lines, offer a fixed line from point to point and are certified to provide data-quality transmissions. They are also relatively expensive because the carrier is dedicating the resource of the line to one customer whether that customer is using the line or not.

Digital Data Service (DDS) lines provide point-to-point synchronous communications at 2.4, 4.8, 9.6, or 56 Kbps. Several telecommunications carriers may provide DDS lines. The carrier guarantees full-duplex bandwidth by setting up a permanent link from each endpoint. Digital lines provide transmissions that are nearly 99 percent error-free. Digital lines are available in several forms, including DDS, T1, T3, T4, and switched 56.

DDS is a digital service and does not use modems. Instead it uses a device known as a CSU/DSU (channel service unit/data service unit), which converts the digital signal generated by the computer into the type of digital signals (bipolar) that are part of the synchronous communications environment. It also contains electronics to protect the service provider’s network.

T1 is the most widely used type of digital line at higher data speeds. This point-to-point transmission uses two pairs of wires to transmit a full-duplex signal at the rate of 1.544 Mbps. They can carry digital voice, data, and video signals. They are quite expensive. T1 channels may be divided among subscribers in 64Kbps increments if the user either cannot afford or doesn’t need the full bandwidth of a T1 line. In some other countries, T1 is not available, but another service, called E1, generally is. E1 carries data at 2.048 Mbps.

T1 uses a technology called multiplexing, or muxing. Several signals are sent to the MUX or multiplexer and fed onto the cable for transmission. At the other end, the signals are de-multiplexed back into their original form.

T1 lines can be divided into 24 separate channels. These channels are sampled 8,000 times each per second. Therefore, T1 can accommodate 24 simultaneous data transmissions over each two-wire pair. Each channel incorporates 8 bits. Because each channel is sampled 8,000 times per second, each of the channels can transmit at 64 Kbps. This data rate is known as DS0. The 1.544 Mbps rate is known as DS1. Ds1 rates can be multiplexed to provide even greater transmission rates known as DS-1C, DS-2, and Ds-4. (Discuss the table on page 588.)

T3 and Fractional T3 leased line service provide voice and data-grade service from 6 Mbps to 45 Mbps. They are the highest-capacity leased lines commonly available. A T3 line can be used to replace several T1 lines.

Switched 56 is simply a circuit-switched version of a 56 Kbps DDS line. The advantage of a Switched 56 line is that it is used on demand, thereby eliminating the cost of a dedicated line. Each computer must be equipped with a CSU/DSU that can dial another Switched 56 site.

In a packet switching network, packets are tagged with a destination address and other information so that each packet can be sent individually over the network. The packets are relayed through stations in the computer network using the best route currently available between the source and destination. Each packet is routed separately, so each packet is sent along a different route. Packet size is kept small, so if there is an error it will be easy to resend the packet. Smaller packets tie up the switches for shorter periods of time. The packets are reassembled in the proper order by the receiving computer. Packet-switching networks are relatively inexpensive because they feature high-speed lines on a per-transaction basis instead of a flat fee.

Virtual circuits are made up of a series of logical connections between the sending computer and the receiving computer. The connection is made after both computers exchange information and agree on communications parameters which establish and maintain the connection. These parameters include the maximum message size and the path that the data will take.

Using switched virtual circuits, (SVCs) the connection between end computers uses a specific route across the network. Network resources are dedicated to the circuit, and the route is maintained until the connection is terminated.

Permanent virtual circuits (PVCs) are similar to leased lines except that the custome pays only for the time that the line is used.

X.25 is a set of protocols incorporated in a packet-switching network, which uses switches circuits, and routes as available to provide the best possible routing at any given time; there are no standard circuits. X.25 incorporates a high level of error checking because it was originally implemented using telephone lines. (Read the rest of the information on page 596.)

Frame relay is an advanced, fast packet, variable-length, digital, packet-switching technology that uses far less error checking than X.25 because of frame relay’s use in a reliable, secure fiber-optic environment. It is a point-to-point system that uses a permanent virtual circuit to transmit frames at the data-link layer. Frame relay networks are much faster than other switching systems because there is no need for frame relay devices to perform fragmentation and reassembly, or to provide best-path routing because the entire path from end-to-end is known.

ATM is an advanced implementation of packet switching that provides high-speed data transmission rates to send fixed-size packets over broadband or baseband LANs or WANs. ATM is part of the broadband integrated services digital network (BISDN). ATM can transmit data at very high speeds of 155 Mbps to 622 Mbps or more.

ATM is a broadband cell relay method that transmits data in 53 byte cells rather than variable-length frames. These cells consist of 48 bytes of application data and 5 bytes of ATM header data. Network equipment can switch, route, and move uniform-sized frames much more quickly than it can move random size frames.

All hardware in an ATM network has to be ATM compatible. To implement ATM on an existing network usually will require extensive equipment replacement. ATM may be used with existing media such as coaxial, twisted-pair, and fiber-optic. However, these traditional network media in their present forms do not support all of ATM’s capabilities. The ATM Forum recommends: FDDI (100 Mbps), Fiber Channel (155 Mbps), OC3 SONET (155 Mbps) or T3 (45 Mbps).

ATM switches are multiport devices that can act like hubs to forward the data from one computer to another within a network, or as router-like devices to forward data at high speeds to remote networks. ATM can use switches as multiplexers to permit several computers to put data on the network simultaneously. (Read ATM considerations to class – Page 602)

ISDN is an inter-LAN digital connectivity specification that accommodates voice, data, and imaging. Basic Rate ISDN divides its available bandwidth. Two of these move data at 64 Kbps, and the third transmits at 16 Kbps. The 64 Kbps channels are known as B channels. The slower 16 Kbps channel is known as the D channel. The D channel carries signaling and link management data. ISDN basic rate desktop service is called 2B+D. A computer connected to an ISDN service can use both B channels together for a combined 128 Kbps data stream. If both ends support compression, much higher throughput can be achieved. ISDN is designed as a dial-up service.

FDDI is a specification that describes a high-speed (100 Mbps) token-passing ring network that uses fiber-optic media. FDDI was designed for high-end computers that did not find enough bandwidth in existing 10 Mbps or 4 Mbps Token Ring architectures. FDDI is limited to a maximum ring length of 100 kilometers (62 miles), so it is not really designed to be a WAN technology.

Networks in high-end environments use FDDI to connect components such as mainframes and minicomputers in a traditional computer room. These are sometimes called back-end networks. FDDI works with backbone networks to which other low-capacity networks can connect. It is not wise to connect all of the data processing equipment in a company to a single LAN because the traffic may overload the network and cause a a failure to the company’s entire data processing operation.

Any operation that requires a high-bandwidth is a good candidate for FDDI.

A computer on an FDDI network can transmit as many frames as it can produce within a given time period before letting the token go. As soon as the computer is through transmitting, it releases the token. Therefore, there may be several frames circulating on the network at once. This explains why FDDI offers higher throughput than a Token Ring network, which only allows one frame at a time to circulate.

FDDI operates over a dual-ring topology, which supports 500 computers over a distance of 100 kilometers (62 miles). Traffic in an FDDI network consists of two similar streams flowing in opposite directions around two counter-rings. One ring is called the primary ring and the other is called the secondary ring.

Traffic usually flows only on the primary ring. If the primary ring fails, FDDI automatically reconfigures the network so the data flows onto the secondary ring in the opposite direction.

Even though the two rings can support up to 1,000 computers and 200 kilometers total distance, the 500 computer, 100 kilometer limit should be observed so that the network will continue to function in the case of a failure of one of the rings.

Computers can connect to one or both FDDI cables in a ring. Those that connect to both cables are known as class A stations, and those that connect to only one ring are called class B stations. If there is a network failure, class A stations can reconfigure the network, while class B stations cannot. (Discuss the FDDI in a Star information on the bottom of page 607.)

SONET is an emerging system that takes advantage of fiber-optic technology. It can transmit data at more than 1 gigabit per second. SONET defines optical carried (OC) levels and electrical equivalent synchronous transport signals (STSs) for the fiber-optic based transmission hierarchy. (Discuss the rest of the page.)

Transmission speeds for SMDS range from 1 Mbps to 34 Mbps with many-to-many connectivity. SMDS uses the same fixed-length cell relay technology as ATM. SMDS does not perform error checking or flow control; that is left up to the sites being connected.