217 Chapter 15: Designing a Network because the data might be static and easy to restore from tape if the disk array is lost. Other applications might require the highest level of data-loss safety possible, with failover servers each having mirrored RAID 1 or RAID 10 arrays and online tape backup systems updating a backup tape every hour or for every transaction. Similarly, some companies might work with data that is so sensitive that they must install the best firewalls, perhaps even two levels of firewalls, and hire full-time professionals dedicated to keeping the data secure. Other companies might be satisfied if they are only reasonably secure. The point is that you must determine how important these issues are to the company for which you are designing the network. Then you can propose different solutions to address these needs and factor them into the rest of your design. Growth and Capacity Planning The final area to consider is the expected growth of the network, particularly if the company expects this growth to be substantial. As mentioned earlier in this chapter, a network designed for a rapidly growing company looks different from one for a slowly growing company, even if both companies start out at the same size. In the former case, you want a design that you can quickly and easily expand without needing to replace much of the existing hardware and software. In the latter case, you can get by with a simpler network design. Consider the impact of growth on the different parts of the network that you’ve already examined (applications, users, and services), because linear growth does not always mean a matching linear impact to the network. Assuming linear growth, the impact to the network might be much lower or much higher than the curve. For example, you saw in Chapter 4 how Ethernet uses a collision detection mechanism to manage network traffic. In that chapter, you also learned that Ethernet scales linearly, but only up to a point. Once the network starts to become saturated, performance begins to drop rapidly because of the chaotic nature of Ethernet’s collision detection scheme. Consider a 10 Mbps Ethernet network transmitting 3 Mbps of traffic. This traffic probably flows smoothly, with few collisions and few retransmissions required. Push the network demand up to 4 or 5 Mbps, however, and its performance grinds to a halt as the network becomes saturated, and you end up with as many collisions and retransmissions as real data. In fact, the total amount of good data flowing over a saturated Ethernet network will be less than the amount flowing over a less-saturated network. You can also find examples where an increase in demand doesn’t cause a corresponding increase in network or server load. For example, the server load for a complex e-mail system might increase only by a small amount if you doubled the number of users, because the system’s overhead generates most of the load. The storage requirements for an accounting system might not double just because you keep twice as much data in it to accommodate the overhead that might consume most of the existing space. Alternatively, that same accounting system might consume four times as much storage space if you double the data storage, because it might have a relatively inefficient indexing scheme. The point is that you need to know how different applications scale 218 Networking: A Beginner’s Guide with increased use. The vendors of the main applications you will use should be able to provide useful data in this regard. TIP Be careful not only to consider how applications behave as they scale, but how they behave as they are scaled in your planned network environment. Different NOSs, network topologies, and client and server computers will all affect how well a particular application can support growth. Meeting Network Needs Once you complete your assessment (by this point, you’re probably sick of the assessment process!), you can then start working on finding ways to meet all the needs you’ve identified. This process is largely holistic and is not worked through by following a series of steps and ending up with a single answer, like an equation. Instead, you should start by mapping out the various parts of the network, considering the three main topics discussed in this section, and “build a picture” of the network design. The design that you create will incorporate all you learned during the assessment process, taking into account your experience and the advice you have received to devise a concrete design that results in an equipment list, specifications, and a configuration. Seeking criticism of your design from other network professionals, who might have valuable experience that you can then factor into your design, is important. No single networking professional has seen and had to cope with all possible design needs, so you want to combine the advice of as many seasoned people as you can. Choosing a Network Type You probably want to start the design by choosing a network type. This should be a relatively straightforward decision, based on the overall bandwidth requirements for the network. For most new networks, you almost certainly will decide to use one of the flavors of Ethernet. Ethernet is by far the most common type of network installed today, and it’s an easy default choice. You also need to decide what level of Ethernet you need. For wiring to the desktop, you should choose 100Base-T. It’s reliable and provides plenty of capacity for most needs. For your network backbone, you can usually use a higher-bandwidth connection, such as 1000Base-T, without incurring too much additional cost. Structuring the Network Next, decide how you plan to structure the network. In other words, how will you arrange and wire the various hubs, switches, and routers that the network needs? This is probably the trickiest part to determine, because it’s hard to predict how much data must flow from any given set of nodes to any other set of nodes. The estimates you have based on your assessment work will help. If you can identify expected heavy traffic 219 Chapter 15: Designing a Network patterns, you should also draw a network schematic with these patterns indicated to help you sort it out. Remember the following tips: N Ethernet’s CDMA/CD collision handling means that an Ethernet network will handle only about one-third of its rated speed. In other words, a 100Base-T segment, which is rated at 100 Mbps, will handle about 33 Mbps of actual data before starting to degrade. N Whenever possible, use “home-run” wiring (in which each network cable runs from each workstation to a single location) for all nodes to a single wiring closet or server room. Doing so enables you to change the network structure more easily (for example, to break segments into smaller segments) as needs change. N Except in the smallest networks, plan on installing a network backbone to which the hubs connect. An Ethernet switch rather than a nonswitching hub should handle the backbone, so each hub constitutes a single segment or collision domain. You still must plan to keep each segment’s traffic below the Ethernet saturation point, but this structure will give you plenty of flexibility to meet this goal. N The physical building might dictate how you structure your network. For example, a building larger than 200 meters (about 600 feet) in any dimension probably means you won’t be able to employ a home-run wiring scheme for all your nodes. This is because twisted-pair Ethernet usually reaches only 100 meters (about 300 feet), which includes routing around building obstructions, patch cables, and other things that make the actual cable distance longer than you might measure on a map of the building. N For multifloor buildings that are too big for a home-run wiring scheme, consider running the backbone vertically from floor to floor, and then have a wiring closet on each floor that contains the switches to service that floor’s nodes. The wiring from the closet on each floor then fans out to each of the nodes on that floor. N Consider running the backbone speed at ten times the hub/desktop network speed. If you’re using 100Base-T hubs to connect to the desktop computers, plan on a 1000Base-T backbone. N Most of the time, most nodes do the majority of their communication to one or two servers on the network. If you are planning department-specific servers or if you can identify similar patterns, make sure that each server is on the same segment as the nodes that it primarily serves. N If your servers tend not to be assigned to support departments and instead support the entire company, make sure that the servers are directly connected to the backbone’s Ethernet switch. 220 Networking: A Beginner’s Guide N If you have any high-bandwidth users, consider keeping them on a segment separate from the rest of the network (if appropriate) and also consider upgrading the speed of that segment to 100 Mbps or 1,000 Mbps if needed. N As you start to implement the network, carefully watch the ratio of collision packets to data packets. If the number of collisions on any segment climbs 5 to 7 percent of the total number of packets, performance is starting to suffer, so you need to investigate the cause and find a way to decrease this ratio. You can usually do so by breaking the segment into smaller pieces or by configuring capable switches into what is called a virtual LAN (VLAN), unless you know of another way to reduce the amount of traffic. Selecting Servers When choosing servers for a network, start by determining which NOS you will use. For PC-centric networks, the decision is usually between Novell NetWare and Windows family of servers. As discussed in Chapter 13, whenever possible, avoid using both, because supporting two NOS systems makes managing the servers much more difficult. You’re better off compromising on a single NOS platform. Next, list the various network services that your servers must provide. You need to look for efficient ways to host these various services on your servers, balancing a number of factors: N All else being equal, using more small servers to host fewer services each is more reliable than using fewer large servers to each host many services. N Conversely, having more small servers increases your chance of having a server fail at any given time. N Using more small servers is more expensive and requires more maintenance than using fewer large servers. N If you plan to use more than one server, consider which services should be redundant on another server or how you plan to deal with the failure of any server. Using your assessment information, you can easily determine how much storage capacity your servers will need. However, it’s much harder to know how capable each server should be in terms of processor power, installed RAM, and other features, such as bus configuration. For these specifications, you need to rely on the advice of the NOS vendor and the manufacturer of the servers that you are considering. Fortunately, both Microsoft and Novell have published tests and recommendations for sizing servers given different service and user loads. Many first-tier server manufacturers also have such data to help you choose an actual server model and its specifications. 221 Chapter 15: Designing a Network Chapter Summary Designing an entire network can be extremely complex. If you are in the enviable position of designing a network, your best bet is to start with the framework described in this chapter and to use other resources to answer specific questions. Many resources are available to help you do this, ranging from books devoted to aspects of network design, server management, network performance tuning, and specific NOS management, to consultants experienced with similar networks and the various vendors you are working with on any planned purchases. In fact, so many resources exist to help you accomplish this job, you may have trouble deciding which advice to follow! Always remember to leave some escape hatches in any network design, so you can respond quickly to new or changed requirements, many of which will occur while you’re finalizing the design. The good news is that if you follow the advice in this chapter and the rest of the book, along with the other resources mentioned, it’s a safe bet you’ll end up with a solid, expandable, maintainable network design that meets the needs of the company and of which you can be proud. . and retransmissions as real data. In fact, the total amount of good data flowing over a saturated Ethernet network will be less than the amount flowing over a less-saturated network. You can. possible, with failover servers each having mirrored RAID 1 or RAID 10 arrays and online tape backup systems updating a backup tape every hour or for every transaction. Similarly, some companies might. the same size. In the former case, you want a design that you can quickly and easily expand without needing to replace much of the existing hardware and software. In the latter case, you can