Q&A: Scaling Application Expertise Helps the Help Desk
Mercury Interactive's new Resolution Center encapsulates expertise for first-tier help desk personnel
If you think the best way to troubleshoot your application performance problems is to bottle the expertise of your second- and third-tier support professionals and make it available to help desk support personnel, then you may want to take a look at service quality management specialist Mercury Interactive’s new Resolution Center.
Application downtime costs money. It’s a fact that IT organizations are increasingly getting wise to, and it’s an issue that this publication has itself explored (see http://info.101com.com/default.asp?id=5564). With this in mind, we spoke with Sunni Gupta, director of resolution and diagnostics with application performance management specialist Mercury Interactive, and Yoav Banin, Mercury Interactive’s director of product marketing.
Both gentlemen discussed the manner in which application performance and availability problems are traditionally addressed in IT environments (hint: it’s a lot like troubleshooting by conference call) and suggested a better way, which empowers help desk support personnel to address even esoteric problems by encapsulating the expertise of second- and third-tier support professionals in searchable troubleshooting guides.
Not surprisingly, Mercury Interactive is shipping a product, Resolution Center, designed to complement its Business Availability Center, which addresses these very issues.
You’ve established a reputation as a service-quality management specialist, but now you’re talking up a new practice, which you call business technology optimization. Can you talk a bit about what this is and why you’re saying it’s important?
Sunni: The problem as we see it is that IT never IT-ed itself. Organizations have been focused on providing IT solutions for manufacturing, for sales, for financial administration—but along the way, IT forgot to IT itself. So Mercury sees a unique market opportunity in order to construct a broad set of solutions such that we can be to the CIO what Siebel may be to some of the other CRM offerings.
What do you mean IT has never “IT-ed” itself?
Sunni: Well, as you know, a lot of IT investments were made in the late 90’s where a lot of organizations were buying tons of stuff without really asking themselves if it was delivering the right (or any) business value. Now, a lot of the priority is shifting to where they want to figure out what’s working and what isn’t—in other words, what’s really delivering business value.
How exactly can they do this?
Yoav: One of the biggest things they’re focusing on is reducing costs by reducing application downtime, or troubleshooting application performance problems. In the past, most organizations focused on gathering bottom-up [information], but what we’re driving with products like our Resolution Center is a top-down approach, starting with the applications, starting with the user—for the first time allow IT to engage the business in a meaningful conversation about what we’re trying to accomplish. So [Resolution Center] is enabling that through service level management and real-time visibility [into the health of applications and systems].
This is important because … there’s a lot of pain with a lot of these large, very complex mission-critical applications that organizations are running, especially because of unplanned downtime. The average cost for some of this unplanned downtime for mission-critical applications is $100,000 an hour, so figuring out the root cause from the top tier all the way down is critical.
What accounts for this cost? Is it mostly lost revenue as a result of application downtime or is it the cost of hiring and supporting help desk personnel to troubleshoot these problems when they arise?
Sunni: If you take a look at the process in the IT organization and break it down into the various tiers, the tier one is the service desk administrators who may be monitoring or taking the calls. The thing is, they’re certainly the cheapest in the organization, and a lot of times, as quickly as possible, these guys are logging the problems without doing any escalating.
Yoav: What’s happening is, where the costs come in, you’re getting these issues, end-user-type issues, application issues, and immediately, all of your experts are called in. You have this bridgeline conference—call it a party line. Everybody’s on the phone, then the incident ping-pong game begins, everybody bringing their own expertise to solve the problem.
If you liken it to emergency medical response, the EMT comes out and provides the first level of care, gets the patient to the hospital to the triage, and then the patient is directed from there. Now imagine that the paramedic comes to the scene, calls to the hospital, orders a busload of specialists to the hospital and they all kind of take turns standing around the guy trying to figure out what’s wrong. That’s the way organizations are doing this today. And this has other effects, too.
Yoav: Application developers, for example, spend anywhere from 40 to 50 percent of their time in support, rather than developing projects. Wouldn’t it be great if we could shift their balance to doing more things that would be helpful for the business rather than plugging holes in the leaking ships? It’s the same for network support people and all of these other guys who are called in on the party line. Wouldn’t it be better if they could be making improvements that add business value instead of just plugging leaks?
Okay. But I’m not sure how another tool, like Resolution Center, for example, is going to help them with this. I saw a recent report from Forrester Research that found that most organizations already have several application performance management tools. Why should another make any difference?
Sunni: Organizations usually have tens or even hundreds of application management tools, and the fundamental difference is that these support mechanisms were architected and designed to support desktops, servers—physical aspects. But once you hit the application, there’s a big gap, and that’s what Resolution Center is specifically designed to address.
Yoav: We provide a lot of things out of the box for those most common application environments, like J2EE, Microsoft .NET. It’s all about scaling application expertise, so you have your tier two [or] tier three individuals or people who know how to solve the problems, how do you enable the rest of the people to benefit from what they know and what they can do? Within Resolution Center, if you think of a decision tree or a Visio diagram, there’s just basically a step-by-step guide on how to automate a problem, we can capture that within Resolution Center. So it’s an easy way for an expert to capture a problem and share it with the front-line people. So for something like Siebel, for example, we’ve spent time to identify those, what are the top 10 or 15 problems in a Siebel environment, and how would you go about troubleshooting them, and we’ve provided this [step-by-step guide] in Resolution Center.
So help desk people can rest easy—a product such as Resolution Center isn’t going to make their jobs irrelevant?
Yoav: The opposite is true, actually. It will empower the front-line support to resolve problems for the first time. A lot of these guys, when they’re getting the calls, they’re typically escalating the problems to the second tier, where it’s a lot more expensive to handle. So we’re empowering the front-line guys so that they can actually resolve a lot of the application problems for the first time. Then for the specialist and the tier-two people, we are also helping them so that [tier one or help desk] is only escalating the problems to the right specialist.
So to sum up, what’s the value add of a product like Resolution Center, and how can it help companies to reduce costs?
Sunni: We automate the setup of gathering that information, so a front-line person can actually run diagnostics against a database or an application server. Because the particulars of that SQL query are already packaged, that solves it. If I’m a DBA and I’ve had the same problem forwarded to me five times, I can just create a runbook [a step-by-step guide to troubleshooting a problem] around those issues. It’s about automating the decision tree, and especially eliminating costly false alerts that are passed on to the tier two and tier three people.