There's a problem with your infrastructure, users are complaining about system slowness. What's your next step?
At this stage, many teams look to the network. They run tests, look for any connectivity issues or broadcast storms, make sure everything's working and consider it a job done. Unfortunately it often isn't. Those same users will often come back complaining of the same issues again or, worse, bypass your IT systems completely and develop their own workarounds. It's not unusual to find employees saving files onto USB keys rather than network drives or using cloud tools rather than company-provisioned applications because it saves them time.
The network does play an important part in this troubleshooting process but it should not be the sole area of focus. Just because the network looks to be performing well on the surface, doesn't mean it's the same for all users.
IT support teams should therefore examine all levels of their infrastructure to ensure the required performance is being delivered. This includes the network, servers and resources, applications and users.
Testing generally does, and should, begin with the network. Organic growth often means the network is not optimised to the level it could be. Is topology correct? Have all switches and users been allocated the bandwidth they require?
It's also important to consider the performance of the network in different areas such as different buildings. Does it perform equally well across the board? Once satisfied the network is delivering the required performance, step one of the testing process is complete.
Servers and Resources
Next, it's time to consider the servers and their resources. In the past, we saw one server tasked with running one application but nowadays, virtualisation is commonplace so it is more usual to see a server running lots of applications. It is therefore more important than ever to ensure that all the necessary resources (connectivity, memory and power) have been provisioned to support these applications.
It is also worth checking the number of virtual machines (VMs) being run via the same physical NIC card. If a VM is known to require particularly large amounts of bandwidth, a NIC card would likely be dedicated to that one VM alone, but where bandwidth demand is smaller, multiple VMs may share the same card. This may be initially viable, but as the number of VMs increases, it is important to be mindful of the performance impact.