|
Network performance is something that affects all of our working lives. The network is literally the central nervous system of a business and therefore monitoring its performance is critical to the efficiency and productivity of any organisation. End user complaints about applications running slowly or delays in receiving email messages are often caused by poor network performance. However, network performance is explicitly linked to the applications that sit on it. As applications become increasingly important to the success of a business, analysing their performance in relation to the network is key to ensuring productivity isn’t harmed by poorly performing IT systems.
Traditionally network performance analysis has focused on utilisation and latency. If a network has been heavily used, IT managers often cite high bandwidth utilisation levels as the reason for poor performance. Although utilisation and latency are important factors they only skim the surface. If network and application performance are to be optimised, IT managers need to drill down and look at how applications are performing and the bandwidth they are using.
In fact, as organisations start to realise that certain applications are critical to business success and that without them thousands if not millions of pounds could be lost, they are beginning to ask for service level agreements (SLAs) from their IT departments. So, for example, if a product ordering application is key to an organisation, business leaders would ask the IT manager to provide guaranteed response times from that application. The IT manager needs to ensure he has the appropriate tools and skills to be able to do this.
IT departments should be looking at the response times which end users experience from applications and examine the bandwidth that each component of an application is using. By looking at these types of metrics as well as the amount of traffic generated by an application, IT managers will be able to deliver more reliable and often shorter end user response times.
There are several methods that IT managers can use to gain this kind of information. Many organisations simply monitor the overall load on the server to ensure that peaks in demand do not have a detrimental effect on the system, and although this provides a basic overview of network traffic, it does not allow organisations to monitor actual end-user experience. Another approach that companies can take is to use a dedicated PC to simulate the steps that users go through when using applications (“transactions”) to provide an insight into user experience. However, this method is synthetic and isn’t always representative of the real user experience.
An optimal solution is to have a method which monitors and manages all transactions in real time and at an individual level, so applications and end user experience can be managed proactively, and before problems develop. As businesses begin to use more and more web-based applications, it is critical that IT managers can quickly pinpoint where problems lie so that they do not waste resources on provisioning new bandwidth for problems which really lie within the coding of the application itself, for example.
With more and more businesses relying on global distributed applications it is important that IT managers not only look at the utilisation of their wide area links but also what it is that is travelling across those links. Furthermore, it is imperative that IT managers use tools which will give them a real insight into the end-user experience of these applications. However, none of this will happen unless IT managers incorporate network and application analysis together and start to look at more detailed metrics that will enable them to deliver a better quality of service to their organisations. |
