Figure 4-1. Increasing effective port density of the LAN Access Level with unmanageable devices
You should consider four main factors when selecting a LAN technology:
Cost efficiency
Installed base
Maintainability
Performance
One of my central assumptions throughout this book is that the network is built for some business reason. It may not directly involve making money, but there must be some benefit to having the network that justifies the expense of building it. Clearly, the benefit is never infinite, so as network designers, we have a responsibility to build a network that meets the requirements for the lowest possible cost.
This problem is particularly important in the selection of network technologies. The classic example is that Token Ring cards for PCs are more expensive than the equivalent Ethernet cards. This fact alone has explained why so many organizations have undergone expensive changes in their LAN infrastructure to use more cost-effective options. As discussed previously, Token Ring has many performance benefits over Ethernet. But if the cost of Ethernet is low enough and the cost of Token Ring is high enough, then you can engineer around the performance benefits to build an Ethernet network that is at least as good as Token Ring, but less expensive. Or, you may decide to spend more money on Token Ring and get better performance.
Similarly, you could get a high-performance network by running Gigabit Ethernet to every desk. But the cost of doing this would be orders of magnitude higher than the same network using Fast Ethernet. There may still be valid business reasons for wanting to build the faster network. However, it is more likely that a hybrid of the two approaches would meet all of the business requirements with a much more attractive budget.
In general, faster technology is more expensive. This is not universally true, however. Fast Ethernet equipment has become nearly ubiquitous, making the cost of building a Fast Ethernet network similar to the cost of building a regular 10Mbps Ethernet. This is even truer of the 4Mbps and 16Mbps Token Ring—it is now difficult to find Token Ring equipment that doesn't support both standards.
The other important cost/performance decision in both Ethernet- and Token Ring-based networks is the granularity of shared and switched segments. The finest granularity network has a switch port for every end device, which has significant performance benefits—particularly because it allows full-duplex operation. However, switch ports are generally more expensive than hub ports. A more cost-effective solution might involve a hybrid network in which some important end devices are directly attached to switch ports, while others are grouped in small numbers on hubs.
Another important economy involves the use of unmanageable Access devices. Small workgroup hubs and switches with no management capabilities are available for remarkably low prices. In the same vein, it is still possible to build an old-fashioned 10Base2 network, using a long piece of coax cable (often called "thin-net"), for almost nothing.
These inexpensive Access options definitely have their place. They may be ideal for the home or small office LAN. They can also be used to increase the effective port density of the network's Access Level by allowing small groups of users to share ports, as shown in Figure 4-1. This figure shows a Distribution Area containing two Distribution switches and three Access switches. Workgroup hubs and workgroup switches are connectd to these Access switches. Some users are connected through the workgroup devices and some are connected directly to the Access switches. Note that I have shown some of these workgroup devices with dual attachments to the Access switches to provide extra redundancy.
This approach works well, but two main structural disadvantages should be considered. First, even if the end devices are able to connect to a workgroup switch at full-duplex Fast Ethernet speeds, they are still constrained by the uplink speed to the Access switch. If the Access switch is also Fast Ethernet, then remember that these end devices must share that link. This option may or may not be acceptable, given the application traffic patterns.
The second disadvantage is the increased probability of failure. The diagram shows that some of the workgroup devices have dual connections to the Access switches, and having these connections is a good way of helping to reduce the net probability of failure. However, workgroup devices are generally not built for the same level of serious use as the chassis switches that I prefer for the Access switches. Specifically, they often have external power supplies of similar quality to those used for low-end consumer electronics.
Augmenting the network's Access Level with workgroup hubs or switches (or passive MAUs in Token Ring networks) is sometimes a reasonable way to reduce costs. Giving up manageability can be dangerous, though, or at least inconvenient. Connecting end devices directly to Access switches allows control over their VLAN membership. Connecting these devices through an intermediate workgroup hub or switch, however, generally means that every device on the workgroup hub or switch must be part of the same VLAN. This requirement affects flexibility.
A more serious problem is the loss of fault management information. An unmanageable workgroup hub or switch cannot tell you when one of the devices misbehaves or when a cable is faulty. It can't tell you when its power supply is overheating. You might be able to get some information about an ill-behaved device somewhere on a workgroup hub by looking at the more complete management information on the Access switch. It can be difficult to narrow down which device is in trouble, though.
Most seriously, if there are problems with one or more devices connected to a workgroup switch, then the only noticeable result will be performance problems for the other devices in that workgroup. The workgroup switch will not pass bad frames to the Access switch,[1] and it can't complain about the bad frames it receives from its end devices. It is possible to have a serious problem that simply will never be seen unless the users are diligent about complaining.
[1] It is customary to use the word "frame" when talking about the Layer 2 view of a chunk of data and the "packet" at Layer 3.
Installed base is another facet of cost effectiveness. The chances are slim that you are building a new network from scratch. In most cases, there is existing equipment, existing applications, servers, and a cable plant. A significantly cheaper alternative network technology may be available. If migrating to that means that you have to absorb a high cost in changing your installed base, then simply staying with the existing technology may be more cost-effective.
For example, a large company may make extensive use of native Token Ring protocols to connect to legacy mainframe equipment. Token Ring equipment is more expensive than Ethernet equipment, but after factoring in the cost of replacing the mainframe, rewriting the applications to use TCP/IP, and changing every end device to use this new application, they probably won't want to make the change.
This is where it is useful to have a long-term strategic information technology vision for the entire organization. If you have a long-term goal to phase out these legacy applications, then you need to build a network that can accommodate a phased-in migration to the target technology. Perhaps you will migrate the Core of the network from Token Ring to Fast and Gigabit Ethernet with TCP/IP routing and use DLSw to tunnel the native Token Ring protocols. Then, when the new servers and applications are available, you can migrate user devices in relatively small groups.
An installed base doesn't need to cripple a network, but it can limit your design options temporarily.
One of the biggest potential hidden costs in a network is maintenance. I have mentioned how using unmanageable workgroup devices in the Access Level of the network can make it harder to find problems. I previously mentioned that the design principle of simplicity makes network maintenance easier. Remember that these are not just annoyance factors for the engineer who gets stuck with the ultimate responsibility for running the network. There are costs are associated with these issues.
The trouble is that quantifying these costs can be extremely difficult. How can you tell, for example, that cutting a particular corner will result in needing an extra staff member to keep the network going? Only experience can tell you what the hidden costs are. In general, since your design goals are centered on reliability, the more corners you cut, the less reliable the results will be. Lower reliability generally translates into higher maintenance costs.
And this topic brings us to performance considerations. You always want to build the fastest and best network you can for the money. Of course, by "best," I mean that the network best fulfills the business application requirements. A brilliant network with unbelievable throughput and low latency is useless if it doesn't support the applications for which it was built.
I mention performance last because it is far too easy to get absorbed in abstract issues of technology improvement. You always have to bear in mind that a network is built for a business reason. It has a budget that is based on how much money this business goal is worth to the organization. If you spend more on building and maintaining the network than it is worth to the organization, either through money saved or revenue earned, then the network actually hurts the organization more than it helps.
Within these limitations, your goal is to build the best network that you can. That also implies that you have to select technology that is appropriate to what you want to accomplish. Part of a LAN may serve an environment where cabling is impossible, so wireless technology could be a natural fit. But wireless technology tends to offer relatively poor bandwidth and latency compared to a similar network built with Fast Ethernet. When selecting appropriate technology, you have to be sensitive to these trade-offs and understand the strengths and weaknesses of the different options available to you.