High-Performance Networking Unleashed

Previous chapterNext chapterContents

- 11 -

Selecting the Right WAN

by Mark Sportack

Wide Area Networks (WAN) are frequently taken for granted. Most users, and even some LAN administrators, don't know what's on the other side of the router connecting them to the WAN. If the WAN was properly designed, implemented, and operated, it is extremely easy to take for granted. It's always there and it always works.

Unfortunately, WANs are rather more complicated than they might appear. Proper WAN design requires an understanding of the various transmission facilities, routers, routing protocols, and the topological arrangement of these technologies. This chapter examines each aspect of WAN design and provides the reader with an understanding of the abilities and limitations of each potential WAN technology and topology.

Internetworking Requirements

Selecting the "right" wide area network (WAN) assumes that some set of criteria exists that the completed WAN's performance will be measured against. If the WAN fails to perform as expected, it obviously isn't the "right" WAN. Therefore, the very first step in selecting the right WAN must be to develop the right selection criteria. Properly chosen, these criteria will guide selection of network technologies, determine the proper size of transmission facilities, and drive the topographical arrangement of the WAN.

Criteria Development

The key issue is developing specific and appropriate criteria. This is a task that is easy to define, yet almost impossible to execute. For example, the potential users of the WAN must be located and identified. A fairly accurate count of them must be made, and correlated to their physical location. So far, so good. The difficult part is estimating their propensity to consume bandwidth. If history can be used to predict future events, users will demand top-of-the-line everything in unlimited quantities--until they get the bill for it. Network planners, contrarily, believe in an obscure law of physics that dictates all available bandwidth will immediately be consumed, regardless of the quantity supplied.

One way to estimate the bandwidth requirements is to identify how the users are currently performing their work. If there are existing networks being used, such as X.25, asynchronous networks, or even modems, they can be invaluable sources of information. They should be monitored to determine

NOTE: Ordinarily, the term bytes should be shunned as imprecise in network planning exercises. In this particular case, bytes is the appropriate term. The average information worker is trained to think in bytes, not octets, so using familiar language will facilitate the data collection effort.

These are vital pieces of information that should form the core of your success criteria as the right WAN will be able to accommodate the projected traffic loads. In combination, these data reveal how much traffic will be put on the WAN and when it will be on the LAN. This is crucial to estimating the bandwidth required across every link of the network.

Other important data that should be determined during this data collection phase is the type of network performance needed. For example, will bulk data transfer constitute the majority of the traffic, or will interactive videoconferencing be the primary application? Is this situation likely to change in the near future? These two particular applications have opposite network performance requirements. Bulk data transfer requires guaranteeing the integrity of the data delivered to its destination, regardless of the time it takes to get it there. Videoconferencing requires the network to deliver packets on time. Damaged packets are as worthless as late packets: They are both discarded. Therefore, it is essential that the performance requirements of the applications be factored into the WAN design.

Another important piece of data that should be determined is the projected aggregate traffic flow. These selection criteria are neither perfect nor complete, but they are an excellent start. Unfortunately, collecting this data won't be quick or easy. In real life, "guesstimates" will likely be substituted for hard facts.

Additionally, if there are existing LANs in use, they must be carefully examined as they will need to be interconnected by the proposed WAN. Important details are

Again, these details should be collected for each and every group of users that will be using the new WAN. Armed with this knowledge, the network planner charged with selecting the right WAN can then consider the two primary aspects of wide area networking: technology and topology. Each of these aspects offers a multitude of options for custom designing the right wide area network for your client base, if you understand their requirements.


The wide area network's technology base includes

Each of these technologies must be examined for their performance capabilities relative to the expected WAN traffic load and performance requirements.

Transmission Facilities

Transmission facilities that will be used to construct the WAN present the richest array of options for the network planner. These facilities come in a variety of sizes and "flavors." For example, point-to-point private lines can range in size from 9.6 kilobits per second (Kbps) to 44.476 megabits per second (Mbps) and beyond. These transmission facilities support a digital stream of data at a fixed and predetermined transmission rate. They can be implemented over a variety of physical media, for example, twisted pair or fiber-optic cabling, and can even support numerous framing formats.

These facilities also vary greatly in the manner that they provide connections. There are two primary types of facilities: circuit switched and packet switched. These two encompass all types of facilities, although technological innovation may be blurring their boundaries somewhat. These technologies are briefly described in this chapter to provide a context for selecting the right WAN.

Circuit-Switched Facilities

Circuit switching is a communications method that creates a switched, dedicated path between two end stations. A good example of a circuit-switched network is the telephone system. Any station can make a dedicated connection to any other station through the central office switches.

Private Lines

The basic circuit-switched transmission facility is a dedicated point-to-point private line. These facilities are leased from a Local Exchange Carrier (LEC) and can be obtained in a variety of forms. They can be either analog or digital, either 1.544Mbps (DS-1) or 44.476Mbps (DS-3) and can deliver service either electrically or optically. They can be also subrated into fractional components, such as 9.6Kbps.

NOTE: The terms DS-1 and DS-3 refer specifically to the CCITT specifications for transmission formats. These terms are often confused and incorrectly used interchangeably with the more familiar T-1 and T-3 terms. The "T" prefix denotes a physical transmission facility, and should only be used to describe the physical facilities.

These circuits provide basic, dedicated bandwidth between two points.

Integrated Services Digital Network (ISDN)

ISDN is a dial-on-demand form of digital circuit-switched technology that can transport voice and data simultaneously over the same physical connection. ISDN can be ordered in either Basic Rate (BRI) or Primary Rate (PRI) interfaces.

The BRI offers 144Kbps in a format known as 2B+D. The 2B refers to two 64Kbps "B" channels that can be bonded together to form one logical connection at 128Kbps. The "D" channel is a 16Kbps control channel used for call setup, take-down, and other control functions.

The PRI is typically delivered over a DS-1 facility at a gross transmission rate of 1.544Mbps. This is usually channeled into twenty-three 64Kbps "B" channels and one 64Kbps "D" channel. Alternatively, higher rate "H" channels of either 384, 1536, and 1920Kbps can be used instead of, or in combination with, the "B" and "D" channels.

NOTE: The 1920Kbps H3 channel is only useable in Europe where the standard transmission rate is 2.048Mbps, instead of the 1.544Mbps that is standard in the US, Canada, and Japan. Attempts to use an H3 channel over a 1.544Mbps transmission facility will result in unuseable channels.

Although ISDN is technically a circuit-switched facility, it can support circuit-switched, packet-switched, and even semi-permanent connections.

Packet-Switched Facilities

Packet-switching facilities feature an internal packet format that is used to encapsulate data to be transported. These packets are then forwarded in a connectionless manner through the commercial packet-switched network (PSN). An example of an old but familiar packet-switched network is X.25.

Frame Relay

Frame Relay, shown in Figure 11.1, is a faster version of X.25 packet switching that features smaller packet sizes and fewer error-checking mechanisms. Frame Relay currently supports only transfer of packets through permanent virtual circuits (PVCs) between the network's endpoint routers. The PVC's endpoints are defined by Data Link Connection Identifiers (DLCIs) and are given a committed information rate (CIR) through the Frame Relay network.

FIGURE 11.1. Frame Relay uses virtual circuits. Frame

Relay requires the establishment of logical pairs of data link connections. These pairs are also given a minimum available quantity of bandwidth, with the option to temporarily "burst" beyond that limit under certain circumstances.

Frame Relay WANs are built by provisioning a point-to-point private line to the nearest central office that provides this service. Much like the central office voice switches that comprise the Public Switched Telephone Network, the Frame Relay switches remain invisible to the user community and their applications.

Frame Relay's primary benefit is that it can reduce the cost of networking locations that are geographically dispersed by minimizing the length of premise-access facilities. These circuits are commercially available at 1.544Mbps, with CIRs used to create logical subrate connections to multiple locations.

Asynchronous Transfer Mode (ATM)

ATM was originally designed as an asynchronous transport mechanism for broadband ISDN. Its low latency and high bit rate, it was speculated, would make it equally ideal for use in local area networks. The subsequent market hype has almost completely cemented its reputation as a LAN technology, to the exclusion of its abilities as a WAN technology.

As a cell-switched WAN technology, ATM is commercially available at either 1.544Mbps (DS-1) or 44.476Mbps (DS-3), although this availability will likely vary geographically.

Customer Premise Equipment (CPE)

Customer Premise Equipment is the physical layer telephony hardware that is required at each customer's premises to terminate the incoming transmission facilities' circuits. Depending upon the type of circuit, CPE can encompass several different devices.

Circuit-switched facilities require the use of Channel Service Units and/or Digital Service Units (CSU/DSU). Customer premise data communications equipment (DCE) are devices that terminate channelized and digital transmission facilities.

Packet-switched facilities require the use of an equivalent device that can assemble and disassemble packets. Such devices are known as PADs.

Premise Edge Vehicles

Routers are typically used to connect LANs to long-haul transmission facilities to create WANs. Although they can also be used as a LAN segmentation device, they are primarily the boundary mechanism that interconnects LANs and WANs.

Internet Addressing

An aspect of the WAN that must be carefully considered is the Internet (that is, Layer 3 of the OSI Reference Model) addressing that will be used. These addresses are used to access and exchange data with hosts on other subnetworks within the WAN. As such, they are a critical component to consider as you select the right WAN for your users.

Theoretically, if your WAN will not be interconnected with the Internet, these addresses could be arbitrarily selected and function perfectly. This should not be done! Use only official, registered addresses in your WAN. This will reduce the workload required to manage the Internet addresses within the WAN, and will prevent duplicate addresses from being assigned.

These addresses will be determined by the routable protocol selected for use within the WAN. Some of the possibilities are: IPv4, IPv6, IPX, and AppleTalk. Each has its own unique addressing scheme. Thus, the choice of protocol determines the possible address hierarchies than can be implemented.

If your WAN requires the interconnection of networks with dissimilar routed protocols, you must have a gateway router at the border of the dissimilar regions. This router must be capable of calculating routes, forwarding route information, and forwarding packets in both protocols.

For more information on the Internet addressing of routable protocols, please refer to Chapter 4, "Internetworking Protocol Stacks."

Routing Protocols

Dynamic routing protocols are used by routers to perform three basic functions:

There are three broad categories of dynamic routing protocols: distance-vector, link-state, and hybrids. Their primary differences lie in the way that they perform the first two of the three aforementioned functions. The only alternative to dynamic routing is static routing.

Distance-Vector Routing

Routing based on distance-vector algorithms, also sometimes called Bellman-Ford algorithms, periodically pass copies of their routing tables to their immediate network neighbors. Each recipient adds a distance vector, that is, their own distance "value," to the table and forwards it on to their immediate neighbors. This process occurs omnidirectionally between immediately neighboring routers.

This step-by-step process results in each router's learning about other routers, and developing a cumulative perspective of network distances. For example, an early distance-vector routing protocol is Routing Information Protocol or RIP. RIP uses two distance metrics for determining the best next path to take for any given packet. These are time, as measured by "ticks" and hop count.

NOTE: Network distances are somewhat euphemistic. They may actually be any of a variety of metrics, and are not limited to physical distances.

The cumulative table is then used to update each router's routing tables. When completed, each router has learned vague information about the distances to networked resources. It does not learn anything specific about other routers, or the network's actual topology.

This approach can, under certain circumstances, actually create routing problems for distance-vector protocols. For example, a failure in the network requires some time for the routers to converge on a new understanding of the network's topology. During the convergence process, the network may be vulnerable to inconsistent routing, and even infinite loops. There are safeguards to contain many of these risks, but the fact remains that the network's performance is at risk during the convergence process. Therefore, older protocols that are slow to converge may not be appropriate for large, complex WANs.

Link-State Routing

Link-state routing algorithms, known cumulatively as Shortest Path First or SPF protocols, maintain a complex database of the network's topology. Unlike distance-vector protocols, link-state protocols develop and maintain a full knowledge of the network's routers, as well as how they interconnect.

This is achieved through the exchange of Link-state Packets (LSPs) with other directly connected routers. Each router that has exchanged LSPs then constructs a topological database using all received LSPs. A Shortest Path First algorithm is used to compute reachability to networked destinations. This information is used to update the routing table. This process is capable of discovering changes in the network topology caused by component failure or network growth. In fact, the LSP exchange is triggered by an event in the network, rather than just running periodically.

Link-state routing has two potential areas for concern. First, during the initial discovery process, it can flood the network's transmission facilities, thereby significantly decreasing the network's ability to transport data. This performance degradation is temporary, but very noticeable.

The second area for concern is that link-state routing is both memory and processor intensive. Routers configured for link-state routing tend to be more expensive because of this.

Hybridized Routing

The last form of routing discipline is hybridization. Although "open," balanced hybrid protocols exist, this form is almost exclusively associated with the proprietary creation of a single company, Cisco Systems, Inc. Hybrid protocols attempt to combine the best aspects of distance-vector and link-state routing protocols, without incurring any of their performance limitations or penalties.

The balanced hybrid routing protocols use distance-vector metrics, but emphasize more accurate metrics than conventional distance-vector protocols. They also converge more rapidly than distance-vector protocols, but avoid the overheads of link-state updates. Balanced hybrids are event driven, rather than periodic, thereby conserving bandwidth for real applications.

Static Routing

A router that is programmed for static routing forwards packets out of predetermined ports. Once this is configured, there is no longer any need for routers to attempt route discovery or even communicate information about routes. Their role is reduced to just forwarding packets.

Static routing is good only for very small networks that have only a single path to any given destination. In such cases, static routing can be the most efficient routing mechanism because it doesn't consume bandwidth trying to discover routes or communicate with other routers.

As networks grow larger, and add redundant paths to destinations, static routing becomes a labor-intensive liability. Any changes in the availability of routers or transmission facilities in the WAN must be manually discovered and programmed in. WANs that feature more complex topologies that offer multiple potential paths, absolutely require dynamic routing. Attempts to use static routing in complex, multipath WANs defeat the purpose of having that route redundancy.


The topology describes the way the transmission facilities are arranged. Numerous topologies are possible, each one offering a slightly different mix of cost, performance, and scalability.

Peer-to-Peer Network Topology

A peer-to-peer network (as shown in Figure 11.2) can be developed using dedicated private lines, or any other transmission facility.

FIGURE 11.2. Peer-to-peer WAN topology.

A peer-to-peer WAN constructed with point-to-point transmission facilities can be a simple way to interconnect a small number of sites.

This topology is often the only feasible solution for WANs that contain a small number of internetworked locations. As each location contains only a single link to the rest of the network, static routing can be used.

Unfortunately, peer-to-peer WANs suffer from two basic limitations. First, they do not scale very well. As additional locations are introduced to the WAN, the number of hops between any given pair of locations is likely to increase. The second limitation of this approach is its inherent vulnerability to component failure. An equipment or facility failure anywhere in a peer-to-peer WAN can split the WAN. Depending upon the actual traffic flows and the type of routing implemented, this can severely disrupt communications in the entire WAN.

Ring Network Topology

A ring topology, shown in Figure 11.3, can be developed fairly easily from a peer-to-peer network by adding one transmission facility and an extra port on two routers. This minor increment in cost provides route redundancy that can offer small networks the opportunity to implement dynamic routing protocols. Dynamic routing can automatically detect and recover from adverse changes in the WAN's operating condition.

FIGURE 11.3. Ring WAN topology.

A ring WAN constructed with point-to-point transmission facilities can be used to interconnect a small number of sites and provide route redundancy at minimal additional cost.

Rings, too, have some basic limitations. First, depending upon the geographic dispersion of the locations, adding an extra transmission facility to complete the ring may be cost prohibitive. In such cases, Frame Relay may be a viable alternative to dedicated leased lines.

A second limitation of rings is that they are not very scalable. Adding new locations to the WAN directly increases the number of hops required to access other locations in the ring. This additive process may also result in having to order new circuits. For example, in Figure 11.3, adding a new location, X, that is in geographical proximity to sites C and D, requires terminating the circuit from C to D. Two new circuits would have to be ordered to preserve the integrity of the ring: one running from C to X, and the other from D to X.

The ring topology, given its limitations, is likely to be of value only in interconnecting very small numbers of locations.

Star Network Topology

A variant of the peer-to-peer topology is the star topology, so named for its shape (see Figure 11.4). This topology can also be constructed using almost any dedicated transmission facility, including Frame Relay and point-to-point private lines.

FIGURE 11.4. Star WAN topology.

A star topology WAN with point-to-point transmission facilities is much more scalable than a peer-to-peer or ring network. Network-connected devices are a maximum of two hops away from each other.

The star topology rectifies the scalability problems of peer-to-peer networks by using a concentrator router to interconnect all the other networked routers. This scalability is available with only a modest increase in the number of routers, router ports, and transmission facilities, compared to a comparably sized peer-to-peer topology. Star topologies may actually be developed with fewer facilities than ring topologies, as Figures 11.3 and 11.4 demonstrate.

The one drawback to this approach is that it creates a single point of failure that can effectively stop all WAN communications. This point of failure, as illustrated in Figure 11.3, is the concentrator node at the center of the star.

Full Mesh

At the opposite end of the reliability spectrum, is the full mesh topology (see Figure 11.5). This topology features the ultimate reliability and fault tolerance. Every networked node is directly connected to every other networked node. Full mesh networks can be built with almost any dedicated transmission facility.

FIGURE 11.5. Full Mesh WAN topology.

A fully meshed WAN topology is readily identified by the complete interconnection of every node with every other node in the network. This approach absolutely minimizes the number of hops between any two network-connected machines, but can be fairly expensive to build and has a finite limit on its scalability.

The reliability of a full mesh network does not come cheaply. To interconnect any given number of nodes requires substantially more transmission facilities and router ports than any other topology. It also has the unfortunate effect of reducing the extensibility of the WAN. Thus, full mesh topologies are more of a utopian ideal with limited practical application.

One application would be to provide interconnectivity for a limited number of routers that require high network availability. Another potential application is to fully mesh just parts of the WAN, like the "backbone" of a multi-tiered WAN, or tightly coupled work centers.

Partial Mesh

A WAN could also be developed with a partial mesh topology. Partial meshes, shown in Figure 11.6, are highly flexible topologies that can take a variety of very different configurations. The best way to describe a partial mesh topology is that the routers are much more tightly coupled than in any of the basic topologies, but are not fully interconnected, as would be the case in a fully meshed network.

FIGURE 11.6. Partial Mesh WAN topology.

A partially meshed WAN topology is readily identified by the almost complete interconnection of every node with every other node in the network.

Partial meshes offer the ability to minimize hops for the bulk of the WAN's users. Unlike fully meshed networks, a partial mesh can reduce the startup and operational expenses by not interconnecting low traffic segments of the WAN. This enables the partial mesh network to be somewhat more scaleable and affordable than a full mesh topology.

Two-Tiered Topology

A two-tiered topology is a modified version of the basic star topology. Rather than a single concentrator router, two or more routers are used. This rectifies the basic vulnerability of the star topology without compromising its efficiency or scalability.

Figure 11.7 presents a WAN with a typical two-tiered topology. The worst-case hop count does increase by one, as a result of the extra concentrator (or backbone) router. However, unlike with the peer-to-peer network presented in Figure 11.3, the hop count is not adversely affected every time a new location is added to the WAN.

FIGURE 11.7. Two-tiered WAN topology.

A two-tiered WAN constructed with dedicated facilities offers improved fault tolerance over the simple star topology without compromising scalability.

This topology can be implemented in a number of minor variations, primarily by manipulating the number of concentrator routers and the manner with which they are interconnected. Having three or more concentrator routers introduced requires the network designer to select a sub-topology for the concentrator tier. These routers can be either fully or partially meshed, or strung together peer to peer.

Regardless of the sub-topology selected, hierarchical, multi-tiered topologies function best when some basic implementation principles are adhered to. First, the concentration layer of routers should be dedicated to their task. That is, they are not used to directly connect user communities. Second, the user premises routers should only internetwork with concentrator nodes and not with each other in a peer-to-peer fashion. Third, the interconnection of user premises routers to concentrator routers should not be made randomly. Some logic should be applied in determining their placement. Depending upon the geographic distribution of the users and the transmission facilities used, it may be prudent to place the concentrator nodes so as to minimize the distances from the user premises.

Given that one or more routers will be dedicated to route aggregation, this topology can be an expensive undertaking. This tends to limit the use of these topologies to larger companies.

Three-Tiered Topology

WANs that need to interconnect a very large number of sites, or are built using smaller routers that can support only a few serial connections, may find the two-tiered architecture insufficiently scaleable. Thus, adding a third tier may well provide the additional scalability they require (see Figure 11.8).

A three-tiered WAN constructed with dedicated facilities offers even greater fault tolerance and scalability than the two-tiered topology.

Three-tiered networks are expensive to build, operate, and maintain. They should be used only for interconnecting very large numbers of locations. Given this, it is foolish to develop a WAN of this magnitude and not fully mesh the uppermost, or backbone, tier of routers.

Hybridized Topology

Hybridization of multiple topologies is useful in larger, more complex networks. It allows you to tailor the WAN to actual traffic patterns, rather than trying to force-fit those patterns into a rigid topological model.

FIGURE 11.8. Three-tiered WAN topology.

Multi-tiered networks, in particular, lend themselves to hybridization. As previously discussed, multi-tiered WAN can be hybridized by fully meshing the backbone tier of routers, as shown in Figure 11.9.

An effective hybrid topology may be developed in a multi-tiered WAN by using a fully meshed topology for the backbone nodes only. This affords a fault-tolerance to the network's backbone and can provide some of the hop-minimization of a full mesh network without experiencing all of its costs or incurring its limitations on scalability.

Fully meshing the backbone of a multi-tiered WAN is just one form of hybridized topology. Other hybrids, too, can be also highly effective. The key is to look for topologies, and sub-topologies, that can be used in combination to satisfy your particular networking requirements.

FIGURE 11.9. Multi-tiered hybrid WAN topology.

Issues with Large Multi-Tiered WANs

Medium and large sized multi-tiered WANs may fall into a common trap: lost focus. Though most common in larger networks, this can strike any company whose network management personnel are evaluated and compensated based on the network's efficiency. Efficiency is a tempting metric because it can be easily measured using superficial techniques. Unfortunately, it is an inappropriate metric that can actually induce inefficiency and increased costs.

Typically, this metric motivates network managers to strive to minimize operational expenses. In a multi-tiered WAN, this can mean using the least expensive method for interconnecting all the user-level nodes to the backbone. Given that most of the transmission facilities available for use in a WAN are priced in a mileage- and bandwidth-sensitive manner, minimizing the mileage of each facility minimizes the cost. Thus, a tiered topology that connects user premises to the geographically closest backbone router represents the most efficient solution.

Wide area networks must be designed to carefully balance costs against performance for all traffic, not just one particular session or location. Since geography seldom, if ever, corresponds to actual traffic patterns, this approach drives systemic inefficiencies. These inefficiencies will be manifested in an increase in the number of hops for the majority of the traffic on the WAN. Consequently, aggregate traffic volumes will be artificially inflated.

Figure 11.10, which shows an example of a three-tiered, point-to-point WAN with actual traffic patterns superimposed, demonstrates the potential inefficiencies of this approach. Access to the concentrator layer is based solely on geography. This minimizes the operational expenses of the network, but results in greater network intensity for any given session.

FIGURE 11.10. Three-tiered, point-to-point WAN.

Thus, the efficient solution from the network management's perspective results in performance penalties for all the traffic on the WAN. The irony is that this inefficiency actually results in a more costly network to operate. Despite this, continued use of the wrong metrics will fail to demonstrate this. In fact, a cost minimization philosophy will continue to demonstrate that the geographically distributed network is the least expensive model, despite its increasing costs.

This is in direct conflict with the original purpose of the WAN! They are not built to provide minimal functionality at the lowest possible cost. Rather, these networks are built to support a company's business processes. As such, the primary metric for evaluating the effectiveness of a WAN should be the degree to which it supports the company's applications and information workers. Therefore, they should be designed to maximize the company's revenue stream, rather than to minimize its expenses.

Once the appropriate network performance metrics are implemented, more effective topologies may be pursued. One such topology is the multi-tiered traffic flow model.

Multi-Tiered Traffic Flow Topology

The solution to the quandary depicted in Figure 11.10 is a multi-tiered topology that is based on actual, aggregate traffic flows. Each router in the concentration layer would be dedicated to one or more groups of users, based on their aggregate traffic patterns. Thus, the topology mirrors the way that traffic flows through the network.

Given that the cost of the vast majority of wide area networking transmission facilities are mileage sensitive, this may be counter-intuitive as it increases the costs of interconnecting user premises to the concentration layer. This sub-optimizes the cost of the premise access facilities. However, the overall cost of running the network will likely decrease because the traffic flow topology, shown in Figure 11.11, minimizes the network intensity of any given session. Please note that this approach is viable only for very large networks that require a tiered topology. It may be possible to implement this topology in a two-tiered WAN, but it really demonstrates the most value in large, three-tiered networks.

FIGURE 11.11. Traffic flow-based WAN.

Figure 11.11 demonstrates a traffic flow approach to WAN design. This approach sub- optimizes access facility costs in favor of maximizing aggregate network performance. A multi-tiered community of interest WAN is built by distributing the premise access facilities based upon traffic flows, not geography.

Proper implementation of this topology will benefit all the WAN's users, not just the ones in well-defined communities. This is because the amount of traffic placed on the backbone is kept to an absolute minimum.


The topologies presented in this chapter can be implemented using almost any combination of networking technologies. Additional complexity can be introduced by examining the subtle differences between vendor brands of each technology product. Selecting the "right" WAN is much more complicated than just picking technologies and a topology. It absolutely must begin with an understanding of the users' collective performance requirements. This is the baseline against which potential technologies and topologies must be evaluated.

The next step in selecting the right WAN requires an understanding of the benefits and limitations of each topology and technology. This must be tempered with an assessment of each one's compatibility with other potential technologies. Other factors, too, must be considered during this process. Embedded base, budget constraints, skill sets, training costs, and even the scalability and expected lifespan of each technology may all affect the selection process.

Last, each technology component must be carefully fitted to the network's topology. For example, using RIP on large, heavily trafficked, multi-tiered WANs would probably be a mismatch noticed by the user community.

Each decision made in the design phase has direct consequences on the functionality of the WAN. These consequences should be evaluated as carefully as the user requirements themselves. For example, an important consideration is how much bandwidth should each physical link in the WAN provide? The consequences of this type of decision are easy to extrapolate. Transmission facilities incur monthly recurring charges that are mileage and/or bandwidth sensitive. Selecting too small a facility may save some money in the short run, but can cripple a company's ability to function. The consequences of selecting an inordinately large facility are usually limited to budget overruns for that expense category. In this case, erring on the side of conservatism dictates rounding bandwidth consumption estimates upward.

The last item to consider as you plan the WAN is the future. A well-designed WAN will not only satisfy its clients on its first day of operation, it will continue to satisfy them long into the future. This requires the network to be robust and flexible enough to accommodate technological change, shifts in aggregate traffic patterns, and growth.

Remember, the WAN exists to facilitate the company's ability to conduct its business. Thus, its success should be measured more by the earnings potential it has created than by the costs it has incurred. With this in mind, study the technological and topological options. Be creative. See what, if any, combinations may be more effective than any homogeneous solution. The right WAN is the one that delivers the performance your user base requires.

Previous chapterNext chapterContents

Macmillan Computer Publishing USA

© Copyright, Macmillan Computer Publishing. All rights reserved.