Application performance monitoring (APM) has become the focal point of IT operations and management. However, managing application performance is no longer the exclusive responsibility of the application team. Today we see network administration teams – that need to prove “it’s not the network” – also stepping up to participate in diagnosing application performance problems.
To accomplish this, there are a number of network-based APM solutions on the market that target network engineers and operations teams. If you’re one of those team members, consider the following as you evaluate new solutions:
1. You don’t always have to start with packets
“Packets don’t lie” is the mantra for most network engineers that are protocol experts. In today's complex distributed enterprise with a multi-tier application environment, given enough time, a protocol expert will find the problem. However, time is always short and the business can never afford to wait.
Therefore, network engineers are constantly looking for ways to shorten their “Mean-Time-To-Know” (MTTK) for where problem domains reside, enabling the problem owner – be it the network, server, application or database groups – to quickly start the troubleshooting process.
A product with information to distinguish between a remote user-problem versus a site issue will help shorten the MTTK. While the network is still an excellent vantage point to see the interactions between users and multi-tiered applications, the initial view into the environment does not have to be at the packet level. It can be equally important to start with a correlated view of the multi-tiered application in action with the users.
2. Don’t use experts until you need them
I recently had a discussion with a network engineer that is part of a team that took over the management of 500 retail store locations for a major chain. The business just started a stores' automation project with new applications rolling out to each location.
Being the resident protocol expert, the engineer is the go-to-guy for any application troubleshooting. This creates a problem: during the rollout process, he would be required to be on the road 100% of the time, tackling triage and deployment issues. In essence, he needs other team members to help him.
A network-based APM solution with built-in streamlined workflow enables other network team members to be productive in isolating the problem domain or decide quickly whether it's a user or site issue. To accomplish this, they leveraged a solution that was the beginning of an improved application problem escalation process that gave the engineer a larger pool of team resources for solving problems - leaving the expert for more challenging problems.
Cross-domain team collaboration
With today’s complex application environment, troubleshooting application performance problems often takes a cross-functional team working together.
Rather than the usual round table finger pointing exercise, APM solutions can provide the initial triage information to isolate the problem domain to the end user, remote site, application, server, network or database, and can speed the root cause analysis process. To facilitate working with IT peers from different disciplines, outside of networking, the information presentation of a solution needs to be usable by the cross-domain team to be truly effective in problem triage.
Network-based APM solutions do not have to take the place of deep-dive products for each IT discipline. Instead, they serve to shorten the MTTK where the problem domain is located – which often takes up the majority of the problem resolution period.
Other than the speeds, feeds, performance, and application support of a solution, consider these collaborative tips and ancillary features when choosing an APM solution. If you do, you and your team could be the next network hero.
Belinda Yung-Rubke is Director of Field Marketing for Visual Networks Systems.