New From NetIQ: Pegasus
End-to-end application performance is the holy grail for new economy network management. While traditional network management applications can tell if application servers are up and network infrastructure is functioning, they rarely offer a glimpse into end user experience.
NetIQ Corp. (www.netiq.com) now offers a monitoring tool designed to give administrators a sense of how well an e-business infrastructure performs for the end user. Pegasus 3.0, the new product from NetIQ, is based on a product by Ganymede, a company purchased by Mission Critical shortly before the merger with NetIQ. Functionality as well as business units have merged - the product integrates with NetIQ’s AppManager application performance tool.
"The question Pegasus asks is ‘Is it the network or is it the application?’" says Jim McQuaid, director of monitoring solutions at NetIQ. Traditional monitoring tools can give performance data about discrete components, server performance, or router throughput. Conversely, applications like AppManager can give indications about the performance of mission critical applications and how well the applications work with the hardware. Pegasus takes an end-to-end view of the network, simulating an end user experience, identifying trouble spots.
"A user may find that the pipe is fine, but the application is slow," McQuaid says. Administrators can use Pegasus to decide which problems are slowing down response time - if the network is fine, administrators can save time and money by immediately focusing on application issues, rather than trying to increase the bandwidth.
Once Pegasus has detected a problem, users can drill down to get a closer look at individual component performance. If it is an application issue, AppManager users can launch the program directly through Pegasus to isolate application-related problems.
Pegasus takes an application-centric view of the network, not surprising from an application management company. "We try to treat the protocol stack as application," McQuaid says. The TCP/IP stack, now an integral part of the operating system, is regarded by Pegasus as an application for networked machines, allowing middleware and front-end web server throughput to be tested along the same lines as host applications.
Another operating system feature that Pegasus monitors as an application is Active Directory. McQuaid says that Microsoft Corp. (www.microsoft.com) has deployed Pegasus internally on over 5,000 servers, partially because of its ability to monitor Active Directory performance.
Network infrastructure, such as hubs, routers, and switches are tested using standard scripts, and integrated into the end-to-end performance window.
NetIQ has bet the barn on Windows NT, focusing most of their products on managing issue related to NT environments. However, while Pegasus runs on an NT console, it can monitor and detect boxes running most operating systems. "All but the most rabid of [administrators] are willing to have an NT box," McQuaid says.
In addition to integration with other NetIQ products, Pegasus 3.0 also boasts a new interface. Administrators may be surprised by the departure from gray numbers and subdued trending information. "We worked with an outside design firm to make the interface snappy," McQuaid says.
Pegasus installation is script and wizard based, allowing some hardware to be detected by the software and giving administrators an intuitive way to tell Pegasus what to look for.
While Pegasus can detect problems across the networked spectrum, McQuaid cautions that simply because a problem is detected does not mean it will be easy to implement a solution. "We’re not going to end the social problems in the IT environment," he says. McQuaid’s experience suggests that coordinating between NT people, Unix people, and infrastructure people can be a huge roadblock to implementing changes in the environment. Pegasus can reduce some of the finger pointing, but inter-departmental animosity may remain. "This really is an organizational issue," McQuaid says.