by Steve Hays
Businesses and organizations today, in order to share information and facilitate better communication in a global marketplace, are unifying their local area network (LAN) resources to form wide area networks (WANs). WANs enable organizations to take advantage of such innovative ideas as cross-functional project teams, distance learning, collaborative workgroups, telecommuting, videoconferencing, multimedia, and access to corporate data.
WANs are growing in popularity, responding to corporate needs for communication and reduced communication costs. The geographic expansion of organizations, increasing numbers of telecommuters, and the growth of client-server and intranet applications have increased the demand for WANs. In constructing wide area internetworks to meet this rising demand, network managers continue to struggle with a variety of issues, including evolving services, emerging applications, and remote users.
In addition, the Telecommunications Act of 1996 is bringing many new challenges and opportunities for any organization with locations in the United States. Long distance carriers, cable providers, and local service providers can all now compete to provide local and long distance service. In the long term, customers will benefit from competitive lower rates and a wider selection of services. In the short term, however, there will be great volatility in the availability and pricing of WAN services.
Many of the potential providers are scrambling to define their service offerings and get them to market. Similar trends also exist on a global scale. During these times of rapid change, organizations need to be able to adapt their networks quickly. Therefore, a key characteristic of remote access and internetworking products is the flexibility to combine WAN services or change from one service to another quickly and easily.
The focus of this chapter is the technologies involved in implementing and supporting a WAN. Information provided in this chapter provides an overview of the steps required to create a WAN, the components that make up a WAN, and the technologies used.
Creation and modification of a corporate WAN is done with the consideration of specific business needs in mind. The following are some of the issues to consider when determining the appropriate WAN architecture:
These issues are defined in this chapter, which functions as an introduction to WAN technology. Designing, developing, and supporting WANs is a specialized science; expertise can come only through education and experience. WANs and their supporting technologies and services are expanding at an unprecedented rate; however, the capable engineers who can plan implement and manage WANs are in short supply.
Building a WAN is a complex activity. A project of this magnitude requires extensive planning, resources, and commitment from the organization. The nature of the organization along with the goals and objectives sought should determine the scope of the project. The process should be iterative, allowing for contingencies and refinements in the technologies that make up the WAN. For more information on these issues, see Chapter 4, "Enterprise Planning and Implementation."
The entire process can be broken down into the following three distinct phases:
See "Building Your Network," [Ch 4]
Again, it is important that this is an iterative process to ensure a long, stable, and secure future for the WAN, and to be sure of its continued value to the organization. As new technologies arrive or the organizational needs change for any number of reasons, there will be a need for continued additions and modifications to the WAN.
To summarize, the design of a WAN consists of the following steps:
1. Identify your requirements.
2. Understand the fundamentals of data communication circuits.
3. Obtain pricing for your area.
4. Evaluate and select the data circuit cost/performance tradeoff that best meets your needs.
5. Understand WAN protocols, and specify organizational standards for them.
By following this straightforward approach, you'll enhance your chance for success.
In order to have program-to-program communication between similar or dissimilar computers, standards must exist. The Open System Interconnection (OSI) reference model is a set of standards that has been developed by the International Standards Organization (ISO). The OSI reference model provides a common ground for manufacturers and service providers to ensure compatibly. Because of the OSI reference model, open systems can be developed with various vendors' equipment and software without the risk of total incompatibility.
The OSI reference model breaks down the communication process into seven layers. Each layer is responsible for providing information and pointers to the next higher layer in the OSI reference model. The application layer makes available network services to software application programs. Each layer describes certain tasks that must be performed by network hardware and software in order for network communication to take place.
Layer 1 through layer 3 support network access, while layer 4 through layer 7 support the communication between the message source and destination. As data is passed through the layers, each layer adds its own information to the packet that defines the configuration the packet is coming from. The data packet is interpreted by the receiver, breaking down each layer and resulting in the reception of the original data. The seven layers are described in the following sections.
Layer 7: Application Layer
This is the layer where applications that facilitate network connectivity reside, for example, FTP, Telnet, and electronic mail, and where network services, including file services, print services, message services, application services, and database services are provided.
Layer 6: Presentation Layer
This is responsible for translating data into formats that can be readily understood by each computer system. The presentation layer concerns itself with translating differences at the bit level, byte level, or character level, and with file syntax. An example of this might be the translating of data formats between two different computer systems that have different data formats. Computers need to agree on the method of identifying the number of bits that equal a whole character and the file syntax that is used by each. The presentation layer also concerns itself with data compression and encryption.
Layer 5: Session Layer
This provides mechanisms that establish, maintain, synchronize, and manage communication between computer systems. The session layer is responsible for the establishment of connection ID numbers, and relies on the transport layer to provide the information that identifies the correct services. The session layer is responsible for the coordination of acknowledgment numbering and re-transmission procedures. It tracks who initially initiated a conversation.
The session layer is responsible for re-establishing a logical communication session between computers should that session end prematurely. The session layer would either resume the interrupted dialog, or initiate another session to the other computer. The session layer also manages a planned connection release of a communication session.
Layer 4: Transport Layer
This provides reliable end-to-end communication by providing service addressing, flow control, datagram segmentation, and end-to-end error checking. The transport layer ensures that packets arrive in one piece, and makes sure that the data is directed to the appropriate service. The transport layer concerns itself with the service addressing, which identifies addresses or ports that point to upper-layer network services.
The transport-level addressing also keeps track of multiple connections or conversations that might occur on a network-attached computer system. The transport layer is also responsible for breaking larger units of data sent down from higher layers into smaller pieces that can be transported across the network. These pieces would then be reassembled at the transport layer of the receiving computer system and passed on to that computer's higher layers.
Layer 3: Network Layer
This is responsible for the internetworking process that needs to occur to reliably send and receive data between networks. It provides logical addressing usually specified in softwareónot hard coded onto network interface cardsóand provides for network routing, flow control, sequencing, and translation. Network-layer flow control monitors network congestion. Network-layer addressing provides addressing that is specific to the logical addressing assigned to a particular network protocol, such as an IP address.
Layer 2: Data-Link Layer
This provides the initial organization of the data bits into a structure called a data-link frame. This frame contains the beginning and end of the frame, source, and destination addresses; a method to ensure that the frame does not contain errors that may have occurred in the course of transmission through the transmission media; and an area to provide some basic administrative functions, such as flow control, frame length calculations, and protocol decisions.
The data-link layer is broken into two logical areas, the Media Access Control (MAC) sublayer, and the Logical Link Control (LLC) sublayer. The MAC sublayer refers to the Media Access Protocol (the way that stations on a network gain access to the media and permission to transmit their data: contention, token passing, polling, and so on), and the physical addressing of stations on the network. This would include the source and destination address sections of the data-link frame.
The LLC sublayer includes portions of the data-link frame responsible for the frame synchronization, flow control, and error checking within the frame. The physical addressing at the data-link layer (used for source and destination addressing in a data-link layer frame) is called a physical address because this address (also called a MAC-layer address) is hard coded into the network interface card in a computer.
This addressing is useful for conveying all the information necessary to direct information to and from computers within a local network. The addressing is usually assigned by the manufacturer of the network interface at the time of manufacture. The first half of the MAC layer address contains hex address information that is unique to each manufacturer; the second half is unique to that individual NIC card.
Layer 1: Physical Layer
This is the foundation of communication between all computer systems. It describes the type of cabling system as the transmission media, the transmission devices that attach to the media, the physical connector specifications, and the electrical or optical signaling characteristics (analog or digital, signal levels, and signal encoding methods). The physical layer also describes the network topology and the distribution of the physical layer transmission media (i.e., in a bus, star, ring, mesh, and so on).
In the OSI reference model, the left stack represents computer number one, the middle two stacks indicate the routed WAN, and the right stack indicates computer number two (see Figure 15.1). The arrows indicate the path taken by data sent from a program on computer number one to a program on computer number two through the routed WAN.
Communications do not always travel through all of the seven layers of the OSI.
NOTE: Data on the receiving computer traverses the OSI in the opposite order it is sent from the transmitting computer.
Communication between two computer systems begins on one system at the highest layer that communication occurs, goes down through to the physical layer, through the transmission media and network, over to the other computer at the physical layer, then up through to the highest layer of communication on the other computer. This communication is usually symmetrical. In the case of this type of communication occurring through a routed WAN, the network portion of the diagram can also extend through the physical and data-link layers and to the network layer.
Not all methods of network communication extend to all the layers of the full OSI reference model. For example, communication between programs on differing hosts using TCP/IP might use all layers, while local Windows for Workgroups communication within the same network using the NetBEUI protocol might have communication occurring only at the data-link layer and physical layer levels.
WANs can be as simple as connecting two distant modems over a standard phone line. Most organizations however, require higher performance than can be achieved through modems. Organizations need to be able to respond to changing communication demands without requiring wholesale revamping of WAN equipment and services. Fortunately, recent network developments have created technologies that offer several choices that meet these needs.
Selecting WAN services is becoming increasingly complex. The number of WAN services to choose from continues to grow, but no single service has emerged as the solution for all situations. Network managers must make constant tradeoffs between cost, performance, and availability. Leased digital circuits, for instance, are widely available but expensive, especially for applications that do not send data continually. Packet services, such as X.25, may be more affordable, but offer lower throughput. Although switched services, such as ISDN, offer higher throughput, they are not yet available in all locations and may be expensive if not used properly.
To add to the dilemma, there can be great variations in the pricing and availability for a given service from region to region. For example, ISDN service may be available in one exchange, but not in the next. If it is available, the tariffs can vary dramatically from one service provider to the next. In the United States, this variability is only likely to increase in the short term with the deregulation of the telecommunication industry. In the long run, increased competition can result in lower prices and greater flexibility, but it is important for companies to pursue WAN strategies that leverage more than one WAN service to retain flexibility as the market continues to change.
Flexibility is thus important in selecting premise equipment, such as routers. Although it might be advantageous to change or reconfigure WAN services frequently to adapt to changing tariffs and minimize WAN charges, the process of reconfiguring, or even replacing, a piece of equipment to handle a different WAN service can be a barrier. If the process is complex and costly, companies may actually find it easier and less expensive to keep their existing configuration and absorb tariff increases rather than change services. It is only when premise equipment allows flexibility that companies can truly design their wide area networks to best meet both their business needs and budgetary requirements.
More than 30 years ago public telephone companies around the world began upgrading their analog services between central offices to digital systems. The phone companies found that digital lines offered more dependable service. Digital services are less sensitive to noise and interference and don't tend to deteriorate with age. Digital systems will either work 100 percent or not at all. Also, digital communication equipment is less expensive than analog. Less circuitry is needed to classify a voltage level as zero or one than to analyze an analog signal for amplitude and frequencies. This enables the phone company to increase capacity for less expense and increase the reliability of its networks.
More recently, developments in fiber optics have once again increased performance and reliability of digital services while cutting the cost of services. Because of the lack of electronic interference, the fiber is more reliable. The data capacity and speed of fiber is orders of magnitude higher than that of copper. For these reasons, fiber is now the media of choice for data transmission. All of the long distance carriers are now equipped with fiber and digital transmitters and receivers.
Many of the Regional Bell Operating Companies (RBOCs) are now upgrading their central office equipment to digital. Also, increase in the use and availability of Integrated Services Digital Networks (ISDN) creates a big push to digital. The ISDN standard, which exists globally, is digital from the ground up.
Although the telephone central office equipment is being replaced by digital equipment, it is still fully compatible with the analog equipment in use today (such as, modems, telephones, and fax machines). The following are some popular network services for WANs:
Each type of access circuit must have a specific type of Data Circuit Terminating Equipment, often also called Data Communications Equipment (DCE) or customer premise equipment, attached to the line in order to be able to transmit data on that circuit.
Service providers usually offer many options as methods of connection through their networks, each with an associated billing agreement. These various options can be summarized as the following three fundamental choices:
The following sections describe some of the digital network services and technologies available today. Where high speed analog leaves off is where the digital services start. For the most part, a 33.6 Kbps modem wouldn't support the WAN traffic of businesses today.
Data Service Units (DSUs) and Channel Service Units (CSUs) are required for any data transmission over all-digital links. DSUs and CSUs in the digital environment are the equivalent of the modem in the analog environment. Most manufactures combine DSUs and CSUs in a single unit because of their complementary relationship. Many of the current analog modems are being built so they can be converted to a DSU and CSU with a simple software upgrade.
The DSU is responsible for the transmission and reception of a signal as well as the buffering of the data and flow control.
The CSU is used to ensure that data terminal equipment (DTE), i.e., computers and network components, do not send signals that could interfere with the telecommunication carrier's network or equipment. The Federal Communication Commission (FCC) requires every digital circuit to be terminated with a CSU.
A permanent virtual circuit (PVC) ensures a connection through a packet network between transmitting and receiving devices. Packet switching uses virtual circuits to make a logical connection to allocate bandwidth on demand between two parties exchanging data. Logic or routing and destination information similar to an address accompanies each packet through the network. When a device sends a packet onto the network, the logical channel number within the packet verifies that the sending device has a PVC connection to the receiving device. PVCs require no call setup or breakdown processes.
A switched virtual call (SVC) is similar to a dial-up call because it requires setup and breakdown processes to take place. The calling device sends a packet over the network known as a call-request packet. This packet contains a logical channel number as well as the address of the device being called. The network uses this address to route the call-request packet to the remote device, usually a Data Communications Equipment (DCE) device that supports the call on the remote side of the connection. A DCE Device establishes, maintains, and terminates sessions on a network.
When the receiving device accepts the call request, it returns a call-accepted packet to the network. The network then sends this packet as a call-connected packet. The channel enters into a data transfer state, establishing an end-to-end virtual circuit.
To conclude the session, one of the devices sends a clear-request packet that is received as a clear-indication packet and then confirmed with a clear-confirm packet. After the call is cleared, the logical channel numbers are made available for another session.
Switched 56 is a digital, time-charged service. As the name implies, the link rate is 56 Kbps with data rates of 150 Kbps to 200 Kbps with compression. Switched 56 is widely available. One good feature of Switched 56 is that it is interoperable with ISDN. That is, a location with Switched 56 service can place calls to and receive calls not only from other locations with Switched 56 service, but also from any locations that have ISDN basic rate service.
As with the office telephone, all you need is the number to call. This enables an organization to use ISDN at all remote sites where it is available and to supplement it with Switched 56 service in other areas, a viable alternative to providing complete interoperability among all remote locations. Switched 56 technology, over the long haul, is expected to be replaced by ISDN equipment.
Availability and economy of T-1 service has increased since it was first offered to the public in 1982. Companies have found that streamlining voice, data, facsimile, and video traffic into a T-1 backbone is more desirable than using the analog networks. T-1s give better control, are easier to troubleshoot and maintain, and are cheaper and faster.
Digital service is based on standard increments of the digital signal (DS). The DS refers to the rate and format of the signal, and the T designation refers to the equipment providing the signals. DS and T are used interchangeably, for example, DS-1 and T-1, or DS-3 and T-3. DS-0 is 64 Kbps, DS-1 is 24 multiplexed DS-0s plus 8 Kbps, for network control overhead, or 1.544 Mbps. With T-1, the 24 channels can individually carry voice or data conversations over the same copper pair required for one analog conversation.
E-1, the European counterpart to the T-1, operates at a rate of 2.048 Mbps. The E-1 is composed of 32 channels at 64 Kbps. Of the 32 channels, 30 are used for voice and data, one for framing, and one for signaling information.
Fractional T-1
In the past, if you didn't need the bandwidth available in a T-1, you had to go to the 56 Kbps line, which didn't provide enough bandwidth in some cases. Fractional T-1 (FT-1) provides the option of leasing only the portion of the T-1 that an application requires, as shown in Figure 15.2. For instance, if an application needs only 384 Kbps of bandwidth, a leased FT-1 with DS-0s 0 through 5 would cover the application and provide significant cost savings over a full T-1.
NOTE: T-1 and FT-1 both enable switching of the DS-0s within the network. This provides flexibility and economy in the configuration of multipoint networks. DS-0s can be switched between destinations within the network using a digital access cross-connect switch (DACS) network.
With Fractional T-1 networks, only the DS-0s needed for the application are leased.
How T-1 and E-1 Work
T-1 and E-1 services can combine voice, data, and video traffic over the same network. Information is encoded using time division multiplexing (TDM) for data and pulse code modulation (PCM) for voice. PCM digitizes voice when carried over digital circuits between the carriers switches. A TDM divides the combined stream of digital information traveling across a link into equal time slots. On the transmitting end, the TDM takes the data from each channel in sequence and places it into a time slot on the aggregate link known as a trunk or backbone. On the receiving end, another TDM receives this aggregate stream of data from the trunk and sorts it back into the original channels (see Figure 15.3).
A single channel is divided into time slots and each transmitting device is assigned at least one of the time slots for its transmission.
Multiplexers are used to merge a number of digital channels into a single link. For instance, by putting a multiplexer on each end of a T-1, all of the 24 channels of data are sent in aggregate. When the data reaches the destination, the aggregate data is separated into its original channels and distributed. Multiplexers are made up of the following four separate elements:
Framing
Framing is the packaging of data so it can be read correctly. T-1 technology was originally based on multiplexing 24 voice channels on two twisted pairs. Each channel carries digitized voice and signaling information in eight-bit bytes, so a frame is formatted consisting of eight bits on 24 channels or 192 bits of data. Added to the framing data is a framing bit in the 193rd position to identify each frame. Each byte is updated 8,000 times per second. This equates to a transmission speed on a T-1 of 192 bits of data, plus one framing bit, times 8,000 seconds for 1.544 Mbps. This description describes what is commonly known as D4 frame and format: frame is the sequence of 193 bits and format is the 24 eight-bit channels or 192 bits.
To understand the term superframe (SF), consider each of the 24 channels on a T-1 as a time slot. As derived from the previous framing discussion, 8,000 framing bits are traveling across the T-1 per second. That's one framing bit every 125 microseconds. The receiving multiplexer looks for a predetermined sequence of bits every 12 frames. It is these 12 frames that make up the SF. There are 12 frames times 193 bits or 2316 bits in an SF.
Within the SF, insignificant bits of data are sacrificed for signaling information necessary for successful completion of the transmission. This process is known as robbed bit signaling.
The problem with the D4 frame and format and the robbed bit signaling is that it was designed for voice streams where it wasn't a problem to replace bits in a voice transmission with signaling bits. However, it becomes a significant problem when data bits need to be removed from a data stream. There are fewer, if any, insignificant bits that can be robbed for the signaling process. This is where the extended superframe (ESF) comes into the picture. ESF increases the size of the SF from 12 to 24 frames, doubling the number of signaling bits available. Also, instead of using 8,000 framing bits, ESF uses only 2,000, leaving 6,000 bits in the 193rd framing bit position for other transmission and error checking functions.
T-1 Fast Packet Technology
Fast packet technology extensively improves the way that information is routed on a T-1 backbone. Fast packet improves the level of support that T-1s can provide for your network. The fast packet multiplexers used with T-1s increases the speed and reliability as well as the efficiency and resiliency of the network. Fast packet can be many times more cost effective for data, voice, and video networks than the traditional TDM used in conventional T-1 networks.
NOTE: Fast packet technology allocates bandwidth as needed. This requires fewer T-1/E-1 circuits to support applications than a standard circuit switched T-1/E-1 network, which divides bandwidth into fixed channel segments.
Remember, a TDM divides the T-1 bandwidth into 24 channels of 64 Kbps feeding voice or data into each of the 24 fixed 64 Kbps channels in the node. In a voice conversation, more than 50 percent of the transmission can be pauses or silence. The classic TDM transfers the silence costing bandwidth, which equates to network performance. Fast packet only allocates bandwidth when necessary. By suppressing silence in voice and idle characters in data, the transmission is many times more efficient.
Fast packet sends data in the same D4 frame that the TDM uses. Instead of allocating one DS-0 per device on the node, fast packet fills the entire packet with data from one channel only with a destination address attached. Fast packet addresses each frame to a single device, so the frame doesn't require any disassembly or re-assembly to pass through intermediate nodes.
Integrated Services Digital Network (ISDN) is a digital switched communication service that is accepted throughout the world. ISDN is capable of sending and receiving voice, data, video, and facsimile over point-to-point digital connections. ISDN was designed digital from the ground up. It represents a standard that is capable of anything from a phone call to WAN connectivity. ISDN is fully compatible with the existing analog services that are available today. ISDN is largely used in telecommuting, videoconferencing, and WAN connectivity.
NOTE: Voice, data, video, fax, and more share one digital link in an ISDN network. One D channel carries the signaling and packet data, while multiple B channels are used for digital access.
ISDN delivers the following types of services:
ISDN Basic Rate Interface
ISDN Basic Rate Interface (BRI) service delivers data on one or two 64-Kbps bearer (B) channels for services of either 64 Kbps or 128 Kbps. Its circuit setup and maintenance signaling is taken out of band to a separate 16-Kbps data (D) channel. It is a switched digital service for dedicated, dial-up applications similar to those that use packet-switched technology. This service is provided by the RBOCs on the same twisted pair that is used by the analog telephone network.
With compression, speeds of up to 400 Kbps are possible over a single 2B+D ISDN line. Users can connect to a variable number of other locations for variable periods of time with sustained throughput comparable to fractional T-1.
ISDN BRI is a popular choice for local access. If there is a flat-rate monthly charge between the locations, ISDN BRI makes less sense if you will be making long-distance telephone calls.
ISDN Primary Rate Interface
ISDN Primary Rate Interface (PRI) service in North America and Japan delivers service through a standard T-1 (1.544 Mbps.) trunk and consists of 23 64-Kbps B channels and one 64 Kbps D channel. In Europe, the service is delivered through an E-1 (2.048 Mbps) and consists of either 30 or 31 B channels and one D channel. PRIs are dedicated trunks that connect to a telephone company's central office. The ISDN PRIs are capable of supporting a large number of voice and data communication. The B channels are used to transmit the physical data while the D channel is used to distinguish ISDN from other digital alternatives on the analog network and tell the network how to handle the B channel data.
Basically, all current telephone and computing systems can be connected to ISDN through a PRI, including PBXs, LANs, WANs, multiplexers, and videoconferencing equipment.
Broadband ISDN
Broadband ISDN (B-ISDN) is the latest ISDN standard that uses fiber optics as a transmission medium. The service supports transmission speeds of greater than 1.55 Mbps and single channel speeds above 64 Kbps. B-ISDN uses ATM as the switching infrastructure.
Asynchronous transfer mode (ATM) has grown out of the need for a worldwide standard to enable interoperability of information, regardless of the network or type of information. ATM has been named as the switching and multiplexing technology for Broadband ISDN. There is an unprecedented level of acceptance throughout the industry of both the technology and the standardization process.
In the past, there have been separate methods used for the transmission of information among users on a LAN versus users on the WAN. This situation has added to the complexity of networking, as users' needs for connectivity expand from the LAN to WAN. ATM is a method of communication that can be used as the basis for both LAN and WAN technologies. In time, as ATM develops, the circuits between LANs and WANs will disappear based on this one standard.
In many instances, separate networks are used to carry voice, data, and video, mostly because these traffic types have different characteristics. Data traffic tends to be "bursty": it doesn't need to communicate for an extended period of time and then sends large quantities of information as fast as possible. Voice and video tend to be more even in the amount of information required, but are very sensitive to the time and order that the information arrives.
ATM does not be require separate networks. ATM is the only standards-based technology that has been designed from the beginning to accommodate the simultaneous transmission of data, voice, and video. ATM provides the following key benefits:
ATM Technology
ATM Technology is based on powerful, yet flexible concepts. When information needs to be transmitted, the sender negotiates a requested path with the network for a connection to the destination. When setting up this connection, the sender specifies the type, speed, and other attributes of the call, which determine the end-to-end quality of service.
Another key concept is that ATM is a switched-based technology. By providing connectivity through a switch instead of a shared bus, the following benefits are provided:
Using ATM, the information to be sent is segmented into a fixed length cell, then transported to and re-assembled at the destination. The ATM cell has a fixed length of 53 bytes. The cell is broken into two main sections: the header and the payload. The 48-byte payload is the portion that carries the actual information: voice, data, or video. The five-byte Header is the addressing mechanism.
ATM System Architecture
The ATM layered architecture enables voice, data, and video to be simultaneously transferred over the network (see Figure 15.4). ATM's implementation is supported through its three lower-level layers, as follows:
The AAL inserts data into and extracts data form the 48-byte payload. The ATM attaches and detaches the 5-byte header and the payload. The PHY converts cells to the appropriate electrical or optical format.
ATM coexists with existing LAN and WAN technologies. ATM specifications are being written to ensure that ATM smoothly integrates numerous existing network technologies at several levels (i.e., B-ISDN, Frame Relay, Ethernet, and TCP/IP).
Frame relay is viewed as one of the most flexible packet switching technologies for efficiently connecting WANs. Frame relay is tightly linked to fast packet switching technology in that it provides an ideal network access technology for connecting data onto a fast packet switching backbone.
Frame relay provides T-1 level speeds using fast packet switching technology for high performance. Frame relay is a WAN technology based on a packet-oriented communication system. Currently, frame relay service is primarily used for local area network interconnections over public or private networks. Other forms of traffic being passed over frame relay include SNA/SDLC, voice, and video.
Frame relay service has been gaining in popularity over the past few years and most major telecommunication carriers offer user interfaces into the packet-switched network. Typically, bandwidth connections range from 56 Kbps to 1.544 Mbps.
Because frame relay is a packet-oriented network service, the user traffic into the network must first be encapsulated inside a frame-relay frame. This encapsulation is performed by a user-to-network interface (UNI) device called a Frame Relay Assembler/Disassembler (FRAD). Most LAN-to-LAN bridges and routers available today can be equipped to provide FRAD capabilities.
At the beginning of each frame of user traffic entering into the frame relay network, the FRAD device places header information that contains the frame address. This information is used by the service providers frame relay switching equipment to route the encapsulated traffic to its destination.
Because packets (frames) of user traffic can be routed through various frame relay network paths, including other carrier's frame relay network and switching equipment, frame relay service specifications exist that define this network-to-network interface (NNI).
Frame relay is a WAN technology that caters to "bursty" traffic. The network guarantees the user a committed information rate (CIR), but permits bursts of data up to the access speed of the connection into the frame relay network. CIR speeds are between 64 Kbs and 1024 Kbs.
X.25 is a low speed packet switching technology that is primarily used for interactive, transaction-oriented applications, such as order entry and credit card verification. It is also ideal for sending e-mail and small files across a WAN. For organizations with a multitude of connections, X.25 technology provides bandwidth between 9.6 Kbps to 64 Kbps at a relatively cheap cost. X.25 is useful for character-based terminal emulation and small volume file transfers. Unfortunately, X.25 does not provide enough support for today's businesses. They demand large file transfers, imaging, video links, client-server technology, and much more.
The Synchronous Optical Network (SONET) provides high bandwidth, high reliability, and manageability, and is well-suited for use as a WAN backbone. SONET was introduced in 1984 by Bell Communications and was quickly accepted by American National Standards Institute (ANSI). SONET rings provide for automatic network backup with 100 percent redundancy so that if there is a point of failure on a fiber ring, service continues on a second ring. SONET rings are currently installed in most major metropolitan areas with larger rings also being installed around multiple major metropolitan areas.
A similar standard, Synchronous Digital Hierarchy (SDH), is established in Europe. For the next few years, the primary use of SONET will be in large telecommunication carrier backbone networks. SONET requires the use of fiber from end to end. Based on light wave technology, speeds range from 50 Mbps to nearly 2,500 Mbps.
SONET is an international standard, fiber optic transmission concept that is used for broadband transport. SONET offers a variety of optical line rates, all of which are multiples of 51.840 Mbps. SONET provides users with the capability to send signals at multigigabit rates over today's single-mode fiber optic telecommunication links, and contains a rich set of operations, administration, and data management capabilities.
Switched Multimegabit Data Service (SMDS) is a high-speed, switched data communication service offered by the local telephone companies that is frequently used to connect WANs. SMDS uses the IEEE standard 802.6 Distributed Queue Dual Bus (DQDB) networking technology and is capable of supporting data rates of up to 45 Mbps. Like ATM, SMDS has a 53-byte cell format. Included in the header of an SMDS cell are control, priority, and error-checking information.
With SMDS, organizations have the flexibility they need for distributed computing and bandwidth-intensive applications. At the same time, because SMDS supports both existing and emerging technologies, it provides the scalability that organizations need to support the applications of the future.
Used to interconnect multiple node LANs and WANs through the public telephone network, SMDS eliminates the need for carrier switches to establish a call path between two points of data transmission. Instead, SMDS access devices pass 53-byte cells to a carrier switch. The switch reads addresses and forwards cells one by one over any available path to the desired endpoint. SMDS addresses ensure that the cells arrive in the right order. With no need for a pre-defined path between devices, data can travel over the least congested routes in an SMDS network, providing faster transmission, increased security, and greater flexibility to add or drop network sites.
Because SMDS is connectionless, it is easy for users to build full-mesh networks in which each site is connected to all other sites. SMDS's three-layered architecture contains the following attributes:
Significant savings are achieved with SMDS because it can deliver the mesh connectivity of dedicated private-line networks with fewer access lines, less terminating equipment, and without the distance-related charges of dedicated networks.
SMDS is the only high-speed, broadband, connectionless data service currently available that offers users a wide range of service features that are generally unavailable on connection-oriented WANs. SMDS offers many features, including the following:
SMDS provides users with the cost effectiveness of a public-switched network; the benefits of fully meshed, wide-area interconnection; and the privacy and control of dedicated, private networks. The key benefits subscribers can realize with SMDS include widespread current availability and increased LAN performance. It provides data management features, flexibility, bandwidth on demand, network security and privacy, multiprotocol support, and technology compatibility.
The LAN-like performance features of SMDS make it a natural fit as a backbone network for seamlessly interconnecting Ethernet, Token-Ring, FDDI, and ATM LANs over extended geographic areas. To connect a LAN to an SMDS network requires only a router and an SMDS-compatible DSU/CSU or SMDS host adapter card. Interface guidelines have been developed by the SMDS Interest Group that enable the service to support the networking protocol architectures found in the leading network environments: TCP/IP, Novell's IPX, AppleTalk, DECNet, SNA, and OSI.
Network managers connect to an SMDS carrier switch via an SNI to a T-1 or T-3 circuit. T-1 SNIs are used to access 1.17 Mbps SMDS offerings, while T-3 SNIs are used to tap into 4, 10, 16, 25, or 34 Mbps offerings. A fractional T-3 circuit can be used to access intermediate-speed SMDS offerings.
With the recently announced low-speed SMDS accessó56 Kbps, 64 Kbps, and increments of Nx56/64 Kbpsósmaller companies and current users of frame-relay technology can take advantage of the SMDS service features. The SMDS Data Exchange Interface (DXI) is used to offer SMDS services at 56 Kbps or 64 Kbps, using the same information formats now employed for on-site connection of SMDS routers to a SMDS DSU/CSU. Because the carrier accepts the data format already produced by the router, a standard DSU/CSU can be used.
SMDS can be used as an alternative to dedicated lines for connecting WANs. The service supports the direct attachment of computing devices for distributed client-server applications, such as database access, file transfer, high-resolution imaging, multimedia mail, and workgroup or collaborative computing.
The features of SMDS, such as call screening, verification, and blocking, also enable SMDS service to function as a virtual private network. This means customers can use SMDS as a public-switched alternative to private networks. Subscribers can either deploy SMDS for full mesh connectivity or use SMDS address screening features to limit transmissions within a closed user group.
Because SMDS is able to coexist with dedicated facilities, it enables customers to create hybrid public/private networks. SMDS also allows for the easy expansion of existing networks because new sites can be quickly added to an SMDS net without totally reconfiguring the network. Additions to an SMDS network only require a simple update to a screening database on the SMDS switch.
The separation of the technology-independent SMDS service layer from the technology-dependent access layers enables SMDS to be supported by many different switching technology platforms and different user-to-network interface technologies. The latest SMDS access interface is being defined using public network, multi-service ATM user-to-network interfaces.
All communications between devices on a network require that the devices agree on the format of the data. Protocols are defined as rules and conventions that govern how devices on a network exchange information. There are a variety of standard protocols to choose from. Each has particular advantages and disadvantages; for example, some are simpler than others, some are more reliable, and some are faster.
The following key functions are required for successful communication across a WAN:
NOTE: It is important for the sake of compatibility, reliability, and performance that you carefully choose and standardize enterprise-wide networking protocols.
One of the most popular choices to implement WAN protocol functions is the TCP/IP protocol. TCP/IP includes IP for network addressing, OSPF for routing, and TCP for end-to-end error checking. These are widely accepted, open protocols that are published as military standards, but available for anyone to implement. See Chapter 9, "Using TCP/IP with Windows NT Server," for more information.
See "Understanding Network Protocols," [Ch 4]
Proprietary protocols developed by private companies may or may not be published. Whether this introduces a limitation on adaptability or evolution of your WAN should be considered. Novell's protocol family, referred to as IPX, includes their proprietary IPX, RIP, and SPX protocols. These are so widely deployed that they are considered "near open."
In addition, choices must be made for client-server protocols. Popular choices are Windows NT and Novell Core Protocol, among others. E-mail, file transfer, encryption standards, and terminal emulation all fit into the upper layers category. You have many choices for implementing WAN protocols. Ideally, an enterprise will standardize on one set of protocols, thus reducing long-term maintenance and administrative costs.
Some common protocols used in LAN networking do not extend to the OSI network layer. These include DEC LAT, NetBIOS, and NetBEUI. Because they have no provisions for network layer addressing and routing, connecting these protocols between different LANs must be done with MAC layer devices, such as LAN bridges or LAN switches. Protocols that operate at the network layer (such as TCP/IP, IPX/SPX, and DECNet) utilize routers to interconnect networks. Most routers have the capability to route routeable protocols and to bridge non-routeable traffic.
Windows NT networks with the domain models are perfect for the needs of a WAN. Windows NT provides for a single logon validation point rather than having to log on to multiple servers throughout the WAN. The Microsoft networking model provides for a single security token validation that gives the user credentials that can take the user anywhere the network is connected, reducing the overhead and complexity of logging on to each server individually.
See "Understanding BackOffice Security," [Ch 4]
Today's wide area networks have the capability of transporting data, voice, and images to far-reaching locations throughout an organization's global enterprise. Managing and expanding these crucial wide-area connections requires highly trained professionals who are able to assess, select, and implement the appropriate wide area services and technologies from an ever-increasing array of options.
Network engineers must have the up-to-date knowledge and skills that enable them to implement, configure, and troubleshoot complex wide area networks. They must know how to make maximum use of protocols, circuits, and tools, as well as plan for migration to advanced implementations and build wide area networks that achieve the optimum balance of cost, security, and performance. For more information on these and related issues, see the following chapters:
© Copyright, Macmillan Computer Publishing. All rights reserved.