When apps are slow, net managers are wrong until proven right

* Why won’t the rest of IT believe it’s not always the network that’s at fault?

In a series of recent newsletters (see the list at the end of this newsletter), we’ve discussed the fact that the Mean Time to Repair (MTTR) application performance issues can be quite lengthy and we described some of the reasons why.

Given the number of responses that we’ve received, it is clear that we struck a cord with a number of readers. One reader responded that the newsletters were relevant to her because her organization had just spent an entire frustrating week troubleshooting one enterprise application. She noted that the root cause of the problem was never isolated but that a reboot of the three relevant Web servers ultimately resolved the problem.

While she was glad that her organization eventually resolved the problem, she expressed understandable concern about the process that took place. The way she described it, the application and server teams insisted the problem was the network, and of course her organization had to prove the network was not the issue before those organizations would consider looking at their components. As we pointed out previously, the amount of time that the network organization has to spend proving that the network is not at fault before the rest of IT gets involved in troubleshooting is one of the factors that lengthen the MTTR application performance issues.

Our reader went on to say that they were running NetQoS’s SuperAgent and Reporter Analyzer, as well as NetVoyant, and that these tools showed that nothing on the network was outside of its normal long term baselines. They also had vendors such as AT&T confirming that they had an error-free environment. In spite of all of that, they still had a very difficult time convincing the other teams that the network was not at fault, even when SuperAgent clearly pointed to the application server tier as the source of the problem.

She concluded by saying that it makes her feel sad and defeated to think that the investment and effort that her organization has made in deploying monitoring tools was being devalued because not one person outside of the network team believes in and takes much consideration of the data. She ended on an optimistic note: “Momentum is slowly moving in our direction and this amazes me as our tools, visibility and transparency are light years beyond what any other group is doing.”

What really struck us about what this reader had to say was certainly not the fact that within her IT organization the network is assumed to be the source of all application performance problems. That is the case all too often. What struck us was that even though her group has far better insight into the causes of potential problems, they are systematically ignored. This is definitely an issue that we will be coming back to.

Further reading: Don’t assume application performance problems are always network-related; What makes application management so hard to do?; NOCs now in charge of application management.

Learn more about this topic

 
Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Related:
Must read: 10 new UI features coming to Windows 10