By Leon Adato.
In the corporate environment, end users and, more worryingly, the occasional IT pro, are the first to point the finger of blame at the network when an application is sluggish, data transfer is too slow or a crucial Voice over IP (VoIP) call drops, all of which can have a wider impact on the bottom line.
Issues arise when the IT department looks to blame the network as a whole, rather than work to identify problems that are caused by an individual application running on the network. Poor design, large content and memory leaks can all cause an application to fail, yet IT departments can be slow to realise this.
Many companies are reliant on applications to drive business-critical processes. At the same time, applications are becoming increasingly complex and difficult to support, which puts additional pressure on the network. So, the question remains, when there’s an issue with application performance, is it the network or is it the application? How do you short-circuit the ‘blame game’ and determine the root-cause of an issue so it can be solved quickly and efficiently?
The first step is to utilize network performance monitoring. This includes application-aware monitoring capabilities and uses deep packet inspection (DPI) and analysis technology to identify mission-critical applications traversing networks. It gives network teams on the front lines a one-stop shop for visibility into both network performance and its impact on application performance and end-user experience. Network performance monitoring will give visibility into whether the problem really is the network, or whether is it an issue with the application itself.
The application delivery chain, or the ‘app stack’ as we call it, is comprised of the application and all the backend IT infrastructure that supports it, including all of the software, middleware and extended infrastructure that the application requires for performance. So, once IT has identified that it’s an application performance issue, the question is where is the problem? And more specifically, why is the application running slowly and how does the IT team fix it? For example, a given application may run across multiple virtual servers and each server will have its own dynamically configured constraints for memory, CPU and storage. Throw in databases and virtual and physical networks into the mix and it’s a very complex troubleshooting scenario.
Solving this problem requires an integrated and unified view of the full application stack for troubleshooting. There are several practical steps IT pros can take to minimise the time they spend identifying and remediating problems and free up more time to proactively address potential issues and threats. Here are some considerations that will help:
- Weigh up performance and financial cost together: for many organizations, a shift to a cloud model is often a purely financial decision. However, an application-centric IT department should consider performance as much as finances when it comes to compute, storage and networking technology, to avoid hassles later on.
- Consolidate the infrastructure: the more partners involved, the longer it takes to troubleshoot problems inside an application. IT pros should try to consolidate as much of the infrastructure as possible under internal management.
- Seek visibility of the infrastructure: take control across the entire stack and ensure complete visibility of the infrastructure is a key component of what a partner or vendor offers. Reliance on partners’ monitoring capabilities means that, ultimately, the end user experience is on their terms.
Getting up-to-date information about the network’s availability and performance should be a top priority for administrators who wish to avoid the 'blame game.' Network performance monitoring is a must-have to enable users to detect, diagnose, and resolve network performance problems and outages before they become an issue.
Furthermore, as the complexity of networks increases and more businesses look to make the move to cloud based applications, breaking down traditional IT siloes and understanding how critical apps are becoming for business can help IT pros make the transition into the full application stack view to truly optimize performance.
The author
Leon Adato is head geek, SolarWinds.
