Paul Caldwell explains why understanding how your business uses its current network is key to planning for a new one.
Applications keep crashing. Everyone is complaining about the lack of bandwidth. Your network seems to be creaking. And you are under pressure to fix all of the problems, in no time at all and with no budget, either. The good news is that this may all be possible.
You begin by gaining a full understanding of your existing network – this is critical before even thinking about replacement or upgrade. Some questions to ask and investigate further are:
- Do you know exactly what equipment you have and their capabilities?
- Are your links and circuits running at capacity and is traffic actually using them?
- Do you have a good understanding of what applications are running across your network?
- Are you prioritising your traffic correctly with quality of service?
Let’s take each in turn.
An important first step is to get a full inventory of your existing network and the devices connected to it. You can then make an informed decision on whether to keep, replace, recover or redeploy equipment to another site. On small networks comprising only a couple of sites, this is fairly easy to accomplish as you can perform a physical audit of the infrastructure, supplemented by remote connection to the routers and switches in order to build a detailed picture and inventory of the network.
However, it becomes a little more difficult when there are lots of sites and complicated infrastructure. When the network is large or complex, using one of the many infrastructure discovery toolsets proves its worth. Most discovery toolsets perform (with a little preparation) a full automatic inventory of the network. The discovery toolsets gathers detailed information from the device and then connects recursively around the infrastructure until all devices that it can find have been completed. Firewalls can block discovery of large sections of a network so keep this in mind. In addition, most discovery software not only finds switches and routers but all devices including firewalls, PCs, servers, printers and IP phones, to name but a few.
In addition to providing a detailed inventory of the equipment components you’ll also be able to accurately pinpoint equipment that has reached end of life and identify any security vulnerabilities with the software running on the network equipment.
Circuits and interswitch links usage
So you’ve got a good handle on all the equipment you have installed throughout the enterprise, now let’s find out about your circuits and interswitch links to see if you’re actually using them and how busy they are. Obtaining detailed statistics on utilisation and error rates on intersite and interswitch links is crucial so you can identify where problems and high-utilisation bottlenecks are occurring. There are lots of tools to help you with this and some are even free.
You’ve now got a fully up-to-date picture of your infrastructure and its components. Where the traffic bottlenecks are and any areas you need to address during any redesign are starting to highlight themselves. But what is causing these bottlenecks? Do you understand all the applications and services that are using your infrastructure? Fancy new client operating systems, such as Windows 7, are communicating constantly on the network whether it’s downloading updates, checking for new email, or broadcasting on the LAN looking for printers and network services. An overall view of the traffic flows traversing the network will help enormously during any redesign and you might well spot traffic that shouldn’t actually be there in the first place.
One recent example was a customer noticing an ongoing slow application response problem during peak business hours. I discovered misconfigured backup software that was transmitting a multi-gigabit file transfer across the network at the same time (sometimes it really is that simple).
There’s a number of ways of discovering the flows traversing the network from using simple IP accounting or netflow on routers and switches, right through to dedicated probes deployed at critical pinch points in the infrastructure producing extremely detailed reports on individual application usage.
Taking the understanding of application flows a little deeper, most networks nowadays utilise some form of quality of service (QoS) policy that should be marking and prioritising critical applications above more general traffic, like email and web access. At this stage, it’s useful to check that the QoS is actually working as originally designed, to verify that the marking is set correctly for each application.
Only when you have completed and gained the complete picture of your existing infrastructure will you be able to decide where to focus any resources for redesign. You’ll also be empowered to make the correct choices on the equipment types, circuit sizes, how traffic will traverse the network, and what the redesigned network will look like in detail. When it’s time to submit the business case to secure and justify the budget for expenditure it will be much easier than the justification of “we need a new network because we’re having these problems”.
What else are you getting out of it? Well you can use the information you have gathered to stabilise all the software images running on the equipment, resolving bugs and patching security issues. You’ve identified equipment that is end of life and has to be replaced anyway. You’ve got all the information to update network management systems. Future troubleshooting of problems will be simplified, since understanding of the infrastructure for support personnel has been improved due to good documentation. And you’ve got all the information you need should you need to redesign the network and you know what to replace and where.
Paul Caldwell is a technical design engineer with BT Ireland. He can be contacted at firstname.lastname@example.org.