Application performance monitoring isn't what it used to be. Once upon a time, the process involved monitoring a two or three-tiered environment including mainframe, database and application server, all located in a single database. Those days are gone.
Now that applications move dynamically from storage through the data center and LAN, and throughout virtualized environments, networking teams are seeking performance monitoring tools that go deeper and offer a wider scope. That can be difficult, considering most organizations are currently working with a patchwork of SAN, LAN and data center monitoring tools. The result is that organizations are literally losing sight of applications as the span across the enterprise network and the Internet.
"For IT shops, the question is how many places do you have to go [looking within elements of the IT infrastructure] to find what you're looking for," says Dennis Martin, president of Demartek, a testing and research firm. The answer -- way more than there used to be.
Application performance monitoring: Identifying the problems
There are a number of barriers to implementing application monitoring tools in converged and virtualized networks. The first issue is in learning how to map symptoms to problems. That's difficult to do in a dynamic infrastructure in which servers are constantly moving. What's more, there's often no single person in the IT organization that understands all the changing parts.
Then there's the battle of integrating a mishmash of tools to manage various parts of the infrastructure, i.e. application monitoring tools from the application vendors, network monitoring tools from the network vendors, server management tools from the server vendors, storage management tools from the storage vendors, virtualization management tools from the virtualization vendors, third-party application monitoring tools, and overarching IT management products from vendors such as CA, HP and IBM.
At the end of the day, "there are more tools than problems," says Sanjay Castelino, vice president of product management at Solarwinds, the maker of Orion products for application management.
Ultimately, the right tool will offer a wider scope of an application's path, and integrate with existing monitoring software. With this, the goal is to enable networking teams to be proactive rather than having to chase data to understand the problem.
One healthcare system's application monitoring tool strategy
SSM Integrated Health Care, a not-for-profit health system with 5,800 physicians and 22,000 employees across Missouri, Wisconsin, Illinois and Oklahoma, turned to Compuware's Vantage products two years ago to monitor application performance of the organization's largest application, Epic's EpicCare EMR (Electronic Medical Records) software suite.
SSM owns 15 hospitals, has a minority interest in another five hospitals, has affiliations with 16 rural hospitals, owns two nursing homes and has a variety of partnerships with physicians, including Dean Health Plan in Wisconsin.
"Management wanted IT to be able to manage Epic from the end-user perspective to see what they were seeing so that when there was an issue, IT would know," says Pat Campbell, SSM senior enterprise management/back administrator based in St. Louis.
Prior to implementing Compuware for performance management in the St. Louis data center, SSM relied on a patchwork of management and monitoring tools, including HP Operations Manager for servers and applications; HP Performance Agent and Performance Manager for servers; CA Spectrum for network management; and a host of other tools to monitor switches, hubs and connections in between devices. SSM also used other application-specific management tools.
The Vantage tool offers visibility across the St. Louis data center, including virtualization on Windows and Unix, as well as an EMC SAN. "We can see whether the problem initiated in the server, the network or the storage," Campbell says, adding that finding the root cause of an issue takes less time. Once a problem is identified, SSM can still use currently owned point tools to go the next step in analysis.
Vantage also helps monitor applications like Kronos, SAP, various clinical apps, the organization's intranet as well as Microsoft Office file servers. But at SSM the journey has just begun. "We want to push out Vantage to additional applications such as Exchange, VoIP, dozens of smaller software apps, and a risk management application, for example," says Campbell. SSM also wants to move the more sophisticated application monitoring tool to its two other key data centers in Oklahoma and Wisconsin.
Application performance monitoring best practices
While tools from companies such as Compuware and SolarWinds can be helpful, there is a series of application monitoring best practices organizations can follow:
- Focus application management efforts first on the organization's mission critical applications.
- Integrating tools, i.e. plug-ins to systems like Tivoli, OpenView, etc., can help control costs and lower the learning curve.
- Consider value versus complexity when it comes to making applications virtual.
- Increase communication and collaboration between IT teams (namely storage, systems and networking).
- Be selective when it comes to buying and implementing new tools. More isn't always better.
This was first published in October 2010