Customers today no longer tolerate outages or poor application performance. Other companies' comparable offerings are a mere click away, so ensuring optimal performance is one of the most important undertakings for keeping customers happy in the digital age.
Traditionally, network and application teams have been siloed, with each focused just on their own goals and their own domains. However, today these two groups must learn how to combine forces to become one super team, ensuring that any troublesome areas of network performance or application performance are identified and tackled before they can even begin to impact end users.
I will present three key focus areas with advice on how to achieve synergy between the network and application teams so an enterprise can ensure a great customer experience from a performance standpoint.
1) Performance issues can arise anywhere
Performance problems can happen within the network, the infrastructure, the application code, or the end user’s environment. That’s an incredibly complex overall environment to monitor and spot any potential problem areas. Network and application performance teams that focus on just their own domains don’t have a holistic view. This tends to lead to a defensive attitude to problem solving where they spend time explaining why the performance issue isn’t within their domain, rather than taking proactive steps to work with another team to find the root cause.
In this new era, it is essential that both the network performance teams and application performance teams break down the siloes between their zones, and start to collaborate on finding those areas of potential performance issues. To aid in this, enterprises should look at how organizational charts can help them to create a unified team through the management structure. It is beginning to become more common that the network and application performance teams report into the same VP.
2) Disparate tools lead to desperation
Over the years both the network groups and application groups have adopted an array of tools to help them in their jobs. Often these point tools are very domain specific and team members need such specialized tooling as they are responsible for both legacy and cutting edge technology. However, these tools have become a double-edged sword. While they may help in one area, the fact that they don’t work with other tools in other areas increases complexity and has different teams looking at different sets of information and analysis.
This exacerbates the problem I discussed above where teams defend their corner, in this case aided by data that can be at odds with the data being presented by the other group. For example, an application performance diagnostic tool may be showing that a problem does not exist within the application code, while the tool used by the network team might be indicating an issue within the application.
As these tools have proliferated throughout the enterprise, teams can become paralyzed, spending inordinate amounts of time trying to debate and reconcile the disparate information from the tools themselves.
While it might be going too far to say “throw all these tools out,” teams must become aware that such disparate tools can make it harder, not easier, to identify overall problems. In summary, adopt a set of tools that work with one another and start at the network level. The network level is the 30,000-foot view. From there you can gain visibility into the applications and adopt application performance tools that go deep into the code. The
DevOps team can then embrace complementary tools that can diagnose how the application is performing on the network infrastructure itself.
Conversely, teams can consider storing all disparate data in a common repository and use this as its starting point or jump point for all inquiries. This way teams have their network, infrastructure, application and end-user experience data in the same place being shared by all parties.
3) The right teamwork and the right instruments equals a great orchestra
Carrying on from the last point about complementary tools, it is of fundamental importance to ensure that the enterprise has a complete end-to-end performance management strategy in place. It is no good if the teams are working well together if there are still blind spots where pesky performance issues may be lying in wait.
Teams must come together and agree that the data they are looking at is uniform across the board. In order to do this it’s important to get the right instrumentation together and agree on the single source of truth when it comes to performance data. To do so, integrate the metrics from the network performance team with the metrics from the
application performance management team.
Adopting these three approaches of teamwork, tooling and end-to-end instrumentation will take your organization’s performance management from being more like a bunch of soloists trying to play their piece the loudest to a symphony orchestra, all playing together from the same score. In this world of increasing digital complexity it is your customers that win, and that’s always good for business.