Pretty much all vendors want to sell you more. Whether it’s hardware, software, or services, vendors want you to buy more. When it comes to buying more hardware, it’s very hard to ignore the Sirens’ call to get the latest technology. I mean, who doesn’t need 10GB to the desktop? You only have 1GB today, but with all of the new technology, you’ll be behind if you don’t upgrade now!
So you order the latest and greatest and feel like you’re doing a good job by future proofing your network. However, many companies make the mistake of buying more before they really need it. Below are three simple things you can do to make sure that you have optimized your existing network assets before you invest in new capacity:
1. Understand real utilization
If you use a Network Management Software (NMS), look at the historical usage and growth of usage over time. Are your users regularly exceeding 80% of their bandwidth? No? Then you probably don’t need to invest in more capacity at this point. Were users at 10% last year but 70% this year? Maybe. Before you take this as proof that you should upgrade, you really need to understand why the growth occurred. If it’s because they are streaming movies all day, it’s probably a bad bet. If you don’t use an NMS or aren’t monitoring end user ports, make sure to take advantage of the utilization information you already have on your switch. For example, with Cisco switches, when you do a show interface you can quickly see the output rate. By default this is 300 seconds (5 minutes). For the purpose of better understanding the real utilization, try changing this to a larger value like 600 seconds.
2. Distribute backplane port traffic
Now, to be fair, sometimes you really do need the new hardware and you have bottlenecks that are hard to identify. One that still gets overlooked is the backplane utilization. For example, on a Cisco 6500 blade, the ports are in groups of 8 ports on a shared bus to a backplane. You can see how busy your backplane is by running the show system command and looking at the Traffic and Peak values. Next to the Peak value you will also see Peak-Time. This is an easy way to see when your network is really under a high load. If you have heavy traffic on one of your groups, you should consider better distributing them to minimize the impact of the backplane oversubscription. This is a great way to squeeze some extra performance out of your existing infrastructure.
3. Snapshot to Evaluate WAN traffic use
For most admins, the LAN isn’t really the problem. The WAN is the more expensive and limited connection. This is where you need to really understand what’s going on. Many companies regularly upgrade their connections to the Internet because when they look at the utilization information, it’s always extremely high. You have probably heard about NetFlow and how everyone wants you to buy a product to collect the flows so you have a long term picture of what’s going on in your network. However, don’t overlook what you already have. If you just need some quick information, many platforms will give you a quick snapshot of what type of traffic is flowing through. For example, the Cisco ADSM is a great tool to get short term information about what exactly is going across the wire. It’s better to investigate a long term collector so get deeper analysis and alerting, but the built in ones will do in a pinch (or an outage).
There are many ways to get the most out of what you have. The key is to understand the technology you already have available to you and really spend the time documenting what’s going on. Don’t wait for the last minute budget crunch or network crunch to try to put something together. Monitor your network bandwidth and resources to understand what normal usage and growth is so you can make the best decisions about when and where to spend your precious upgrade dollars.
About the author: This post was submitted by Mav Turner, Group Product Manager for Solar Winds.