In-Depth

Q&A: Automating Your Application Infrastructure

A new survey reveals how IT is trying to automate the application infrastructure.

• By James E. Powell
• 11/11/2008

In July, mValent conducted a survey of IT teams in Fortune 1000 companies, asking about their priorities and challenges in managing application infrastructure to support mission-critical applications and services. We spoke with Jim Hickey, chief marketing officer at mValent, about what 286 senior IT professionals had to say and what their answers mean for your own shop.

Enterprise Systems: In your July survey, the top-ranked initiative was "ensuring high availability for applications and/or business services." Why does that head the list of concerns, ahead of deploying/maintaining virtual server environments and implementing SOA, two hot-button issues?

Jim Hickey: I think there are a couple of factors at work here. Often, we tend to focus on technology trends and lose sight of the fact that IT exists to serve the business- -- to help grow revenue, to streamline operations and to cut costs. Our survey highlighted the fact that downtime costs real money- -- over $100k an hour according to a third of the respondents. The drive for competitive advantage in the marketplace also means that applications must stay up and running 24/7/365.

Keeping all these apps running means you need effective application configuration management. What gets in IT's way?

I think there are three things in play here. Number one is change- -- change is the enemy of a smooth-running application environment and change rates for configuration settings are going through the roof.

Then you have a huge volume of applications in a large company- -- Forrester says hundreds to over a thousand typically.

Finally, you have a large number of people involved in making these changes; more than 10 people for most companies in the Fortune 1000. So to net it out: lots of applications plus lots of change plus lots of people equals lots of problems.

That all adds up to a lot of complexity.

That's right. The number one issue is complexity and that it has at least two components. One is solely the complexity involved in managing configurations within any given application whether these are n-tier, distributed applications or whether they are legacy or custom applications. While they are complex on their own, the management problems rise geometrically when you factor in the large number of instances and environments that IT teams need to manage.

Virtualization is a tremendously popular technology, especially in light of "green" initiatives, lowered costs, and ease of implementation. However, the survey points out a downside: rising complexity.

This was a real shocker- -- at least to us as a vendor. Thirty-seven percent of respondents said virtualization made their IT environments less complex. But the rest- -- 63 percent either said that virtualization made things more complex (27 percent), that it made no difference (13 percent) or that they just didn't know (23 percent). So almost two-thirds of the group could not give an unqualified "'thumbs-up"' to virtualization- -- only one of the biggest technology trends happening. That sure was interesting.

Discuss what the survey found out about how IT is trying to automate the application infrastructure. What's driving IT's adoption of automation tools, and what are the benefits IT expects?

Here are three things that jumped off the page: First, eliminating configuration differences between environments (like between QA, Staging, and Production). Everyone knows that the code is the same in the three environments, so why does it work well in once place but not the others? Figuring that out and finding those needles in the haystacks of configuration settings is a major time sink and delay.

Second, we see a continuing need to improve productivity. Everyone needs to do more with less, so investing in automation tools for the application infrastructure is making a lot of sense to IT execs, particularly when configuration issues are at the heart of some many outages.

Finally, respondents said they need faster troubleshooting so that applications can be back on line faster. That, of course, leads to overall revenue and productivity gains for the company.

Are there other solutions besides automation tools that you recommend an IT organization consider?

Yes, definitely. In addition to some of the automation tools that we've spoken about, I think customers should also think about solutions that focus on their internal processes. These could be things like work flow management or process automation tools.

Another area to look into is project or program management software. Both sets of solutions that I have mentioned also address the question of how to make your staff more productive -- but they do so from a different angle.

What surprises did you find in the survey results?

That virtualization may not be all that it seems. When 63 percent of people can't give a positive endorsement of what is seen as a key technology trend, I think that's a real eye-opener. Instead of completely solving a problem, it may have shifted the problem from hardware cost and power consumption to increased challenges in systems management in both labor costs and overall complexity.

What best practices can you suggest for IT teams looking to automate their application infrastructure?

Focus on people and process. Start looking there and then pick tools that either support those people and processes or induce needed change in your internal processes. I think that's an often overlooked point. Simply introducing technology with the intent of solving some of these problems will not get the job done.

In reality, a company's most critical resource is its people. They can either make or break any technology initiative. Secondarily the work processes and organization around the people also affect the success of these initiatives, so you need to select the tools not in isolation but in the context of your people and process.