Rolling Out a New App? How to Avoid 3 Performance Pitfalls
May 09, 2012

Denise Dubie
CA Technologies

Share this

New applications bring the promise of increased productivity, happier end users and better business benefits. But if the new app stalls on the network, IT managers will only be hearing about the problems the technology causes.

Casey Louisiana, sales engineer at Network Instruments (a CA Technologies partner), meets with customers regularly as they work to deploy applications and keep the network running smoothly. He says there are a few common situations that happen frequently, which could be easily avoided. Here he offers a few fixes for these often overlooked issues that can crop up when IT is rolling out a new application.

1. Configuring Quality of Service

Often when IT departments are looking to add, say, voice over IP, to their application mix quality of service sometimes becomes an afterthought. With different management metrics for traffic (for instance, voice and data behave differently) and various prioritizations for some application traffic, IT managers don’t always configure routers, switches, servers and the like to optimally handle the traffic.

“Networking people generally know that they have to pave the road for an application to perform on the network, but sometimes the application folks don’t realize certain apps behave differently,” Louisiana says. “Configuring quality of service correctly can put the appropriate controls in place before deployment, and after the fact, reconfiguring devices could help. In some cases, you have to get down to the code level.”

2. Right-Sizing Applications

Even in today’s fast-paced environments, it’s not good practice to roll out an application without first testing it. Still Network Instruments’ Louisiana explains that some test environments don’t adequately replicate a real-world network, causing application parameters to fall short and performance to lag.

That’s why developers should model situations in which the application will have to perform. This type of “what, if,” service modeling (or service virtualization) can help application developers and IT managers better understand how an application will perform alongside other traffic types, under heavy load and with additional users. “It’s a good practice to apply conditions to better understand what the network and application can handle,” Louisiana says.

3. Assessing App Behavior

Even apps can’t communicate properly at times, according to Louisiana. Technologies from different vendors that is said to integrate or work seamlessly might not always deliver on the promise, causing applications to make requests in what is essentially a foreign language to some devices. Understanding application behavior and knowing the dependencies among underlying devices can go a long way to improving performance.

“Sometimes an application could be asking for ‘A’ and getting the answer ‘3’ from the device. They don’t always work together,” Louisiana says. “Identifying the unlike technologies and how they are designed can help improve application performance.”

Denise Dubie is New Media Principal in CA Technologies Thought Leadership Group.

Related Links:

Share this