Keep your Network on its Toes

Let’s face it. There’s no such thing as a perfect network, else there’d be no role for a network administrator and support staff. You would put a network in place, put your users on it, and be happy about it for the rest of your life. But this is a real world, where networks do get stressed out, become sick, and give performance bottlenecks. To keep it healthy, you have to find ways and means of constantly monitoring and maintaining your network. Let’s see what it takes to do that.

When optimizing networks, there will be some limitations you’ll have to work with. These will be factors beyond your control.

For instance, you’ll always have users complaining about things not working no matter how good your network is. They’ll cry blue murder if their problem isn’t resolved within minutes of their complaint. You’ll also have smart-alecky users, who will install unwanted things on their systems, which will create unnecessary trouble. You will always get all sorts of nasty viruses and Trojans on your network, which you’ll have to identify and remove. You will also always have hackers and crackers trying to get into your network to steal data. Another more serious limitation today is the economic slowdown, which is causing a lot of companies to do budget cuts. If your company is unwilling to invest in IT infrastructure, then you’ll have to work with what you have and work out ways of using it optimally. Knowing the limitations will help you put your task in the right framework.

Know thy network



The journey to optimizing your network starts from documenting what’s on it. I know it’s the most boring job in the world, but it’s just one of those things you have to put up with. You should have complete and updated documents of everything on your network right from the cabling infrastructure, hubs, switches, etc, to all the different servers (mail, file and print, Internet, backup, etc), client configurations, printers, and even the applications. If it’s a small network, then perhaps a simple spreadsheet listing the various items along with their details might do. A slightly larger network might require a small database to keep track of the equipment. For very large networks, you will need additional software, like MS Visio, to create your network diagrams, and network-management software to manage your network infrastructure. These will help you track and pinpoint problems quickly and easily. You should also maintain some sort of troubleshooting guide containing all the commonly faced problems and their solutions. Having this information handy and easily accessible will help speed up troubleshooting. For instance, if a user complainins of not being able to login, you should know before hand his client machine’s configuration, location, and applications running on it. Check your records for other occurrences of this problem and how they were resolved. Knowing this, you may be able to solve the user’s problem over the phone rather than going there yourself.

Besides documentation, you need network-monitoring tools to measure the most important characteristics of your network. Some of these are the network throughput, Internet bandwidth, e-mail/messaging system, and network packet flow. These will help you determine whether your current network policies are fine or need modification, whether it’s time to upgrade your network infrastructure, etc.

The different types of network monitoring tools have been covered in the article Choosing the Right Network Tool on page 84. Whatever tools you run, you should run them periodically so that you have data over a sufficiently long period of time. This will help you separate the normal patterns from the irregularities. For instance, suppose your normal network usage during peak hours is 3 Mbps, which increases to 4 Mbps, then it should immediately start ringing bills. Try and find out which devices are generating this extra traffic.

Sniff out trouble



The number of complexities involved in network troubleshooting and optimization varies according to the network size. For instance, optimizing a small peer-to-peer workgroup having just a bunch of Windows 9x machines would be far easier than a large network consisting of hundreds of nodes. That’s because they would both differ in many ways such as the core structure, topology and number of devices. So the best way to sniff out trouble is to first identify the factors crucial to your network’s well being, and determine how you’ll monitor them. Which are the areas that could bring your entire network down if not checked on time? Which ones will severely hamper work on your network? It’s by asking these questions that you’ll know what to monitor, and come out with a solution based on the results. For instance, some of these could be your network’s main backbone, Internet connectivity, and messaging systems.

When we talk of a network backbone, several devices come into the picture. These include the hubs, switches, and routers on your network. While these devices are pretty sturdy and don’t really go down that easily, you generally have to look for traffic patterns across all of them. For instance, if certain nodes on your network are generating excessive traffic, you have to identify them and decide whether they should be placed on a separate segment or put inside a VLAN (Virtual LAN).

Most medium-sized organizations with a network today would have or would be considering opting for a leased-line connection to the Internet. As bandwidth is still not very competitive at such a level, it has to be managed properly. Therefore, what’s needed is a suitable policy that will help you control and manage the bandwidth effectively. Points of concern here would be controlling access to websites, chat, and

e-mail.

Once you’ve identified the areas of concern, the rest is a matter of trying to prevent major problems from occurring. Remember that network troubleshooting and optimization requires more than just technical knowledge. It’s more of an art of being able to handle situations.

Anil Chopra