Have your business applications given you performance headaches at some point during their lifecycle? Maybe you resolved the problems and the application was able to keep servicing your customers and performing its business function for you.
But no doubt there were other times when you tried to solve the performance problems internally, but sadly-and expensively-you didn’t succeed. To keep future hassles away, you probably (a) added more hardware or, worse yet, (b) started a server recycle program to minimize the damage to your production.
As a result, the performance is still sub-optimum, your company is losing money, and your customer satisfaction keeps sinking. To get your application back on track, you’ve decided it’s time to bring in a professional performance consultant.
You’ve made a smart decision. But now you face the challenge of finding the right person. You must find someone who can quickly identify the performance problems, resolve them, and implement a plan to bring stability to the application.
But that’s just for starters. You also need someone who’ll work well with your team and can put standard operating procedures in place to stabilize the current environment and prevent future problems. Plus, this person must be able to transition the best practices to your team clearly and effectively.
As you can imagine, it’s tough to find all these qualifications in one person. Performance consultants are not all created equal. When the time comes to look for your professional performance consultant, here are the 10 key things to keep in mind:
1) SOLVABILITY: The performance troubles you’re facing are common, and cost many companies money. The good news is these problems are solvable. But companies often have a consultant look at an application and he or she decides the problems are just too complex or aren’t solvable.
For example, one company had a software pricing engine that was critical to the business: if the pricing engine was down, the company lost revenue. The CIO admitted that to remedy these problems, they’d have to add more servers. The standing architect said he’d need 240 new servers to handle the load volume. advice
I came in, did some testing and found that the performance troubles stemmed from the amount of memory the current application was using. My options in this case were limited: I couldn’t rewrite code or change architecture, but I was able to change the Java Virtual Machine for the poorly performing application. This alternative JVM was more forgiving on memory consumption. With new settings, the company was able to scale back from buying the 240 new servers to only 10. Imagine the cost savings!