Group actively monitors application performance end-to-end
American Heart Association is using various application and Web management tools to optimize its Cisco network.
American Heart Association keeps tabs on WAN performancePodcast: At American Heart Association good network management is good security
“We know that every dollar of funding research saves a certain amount of lives,” says Josh Hinkle, manager of network management and security at the American Heart Association and a featured speaker at Network World’s recent IT Roadmap event in Dallas. “So we look at the bottom line for every technology spend and say, ‘How is this going to help us raise more money?’ From a network standpoint, if we can provide a higher uptime on Web applications and business applications, and the Web sites are more available, people are more apt to do business with us, creating more revenue. And save more lives.”
Providing better uptime for Web applications means extending focus beyond the network and homing in on the entire user experience, from browser to application to server to network and back. “The problem isn’t always the network,” Hinkle says. “And we needed a way to show that.”
Round-trip monitoring
After a recent IT reorganization within the American Heart Association, in which 11 field IT units were combined with the main headquarters IT division to form a centralized group, the need for better network oversight and efficiency came to the fore.
“We originally were looking at overall security issues, but when we started doing that, one of the biggest things that jumped out at us was that security starts with network efficiency,” Hinkle says. “Because if your network isn’t efficient, you can’t identify some of the more common and simple attacks like DDOS and viruses breaking loose, because you don’t know when the traffic trending doesn’t fit into your normal patterns.”
The organization had some monitoring tools in place, such as Websense Enterprise policy-based Internet filtering software and Fluke Networks’ Visual Uptime Select for network performance, but it had no way to do real-time application performance-monitoring and -reporting. “We could do weekly or monthly reports on network usage, but we needed something that would alert us and provide in-depth reports in five-, 10- or 15-minute increments,” Hinkle says.
To fill in that blank, Hinkle opted for NetQoS Performance Suite, which ties in closely with the American Heart Association’s Cisco infrastructure. It leverages the traffic statistics gathered by the organization’s Cisco IOS NetFlow- and IP Flow Information Export (IPFIX)-enabled routers and switches to provide real-time visibility into network traffic and application performance across the enterprise.
“Because it works with NetFlow, it allows us to really measure every application’s round-trip times,” Hinkle says. “So it lets us break everything down and see when the degradation happens at a server, at the application or the network. And that’s what really took us to the next level. Not only did it show that information, but it alerted on that information in near real time.”
Proactive prescription
Although the results were not surprising to Hinkle and his network group, they made for fascinating reading on the part of the various application and server groups within the association.
“The worst way to manage your network is to make your decisions based on assumptions; you need to have the true data to work with,” Hinkle says. “NetQoS helped with that by letting us confirm that sometimes, the data transport isn’t the problem. Sometimes the application isn’t performing, or the server is overutilized.”
Hinkle says the NetQoS reports are very visual and easy to interpret, even for network novices. Once an application or server group begins receiving the NetQoS-generated reports, it helps head-off a lot of discussion and lets the whole IT organization become more proactive about finding and fixing problems at their source.
He offers the example of a recent trouble ticket he received. “One of our Web apps wasn’t working, and everyone’s first thought was that it was a network problem,” he says, noting that a quick check of NetQoS and his other tools showed that the network was humming along nicely. “Turns out that it was actually because the Java embedded in somebody’s browser wasn’t working – it was an application problem.”
Even so, Hinkle stresses that to meet the end goal of network efficiency and critical application uptime his network group needs to be a part of the solution, even when the problem isn’t network-oriented.
“We’re not trying to throw our business partners under the bus,” he says. “When failures happen, we work to help our partners fix them, and that’s when people start seeing the efficiency and we start having better overall application performance.”
Hinkle says his group is in the midst of getting every critical application into the monitoring mix, which he says is a fairly simple task that requires only collecting server, port and IP address information on each application. “We enter that info into NetQoS and instantly, it starts collecting data,” he says.
Once all of the applications are being monitored in real time, Hinkle says his group will focus on increasing overall network efficiency.
“We’ll begin focusing on what we can do to better transport data,” he says. “And that will lead us down the road to some sort of WAN acceleration or data shaping — things along those lines to better manage it once we know it’s tweaked to where it needs to be.”
Hinkle says the American Heart Association is already seeing results from its more active attack on application-performance problems.
“We’re finding that a lot of the people we serve aren’t even in the U.S. – many of our researchers are from other countries, Europe, Japan, Canada,” he says. “And we can really see the efficiency of the Web applications driving back through the more varied people requesting grants. The base is broadening because the applications are more efficient.”
Cummings is a freelance writer in North Andover, Mass. She can be reached at jocummings@comcast.net.
