Network Instruments’ Observer Combines Physical, Virtual Visibility

Tracks network, application performance in both virtual and physical environments

• By James E. Powell
• 03/23/2009

Network Instruments has released an upgrade to its Observer monitoring platform that adds support for virtualized environments. Observer provides “visibility and in-depth analysis of application performance and traffic,” the company announced in a press release today.

Charles Thompson, Product Manager at Network Instruments told Enterprise Strategies, that in conversations with his customers, most report that they are unable to access data streams via purpose-built devices (such as analyzers or IDS), and that makes for blind spots when they try to analyze internal VM host traffic. “Their existing VM monitoring lacks in-depth performance metrics, and they must choose between views of the physical world or the virtual world. What they really need to do is combine the two.”

As Thompson explained, “When you use SNMP to poll a physical machine for information about its CPU load, that’s a relevant and important statistic because it’s telling you what that CPU is actually doing. However, when you poll a virtual machine, it has a simulated CPU, so it’s simulating CPU cycles, there are delays inside the communication path that don’t exist in the physical world, and there’s a relevance and a meaning that gets lost when you ask what a simulated CPU is doing.”

How does the company get into the virtual machine to accurately get the right data?

In a Network Instruments release, the company says that “a user installs a Network Instruments software probe into their VM host environment, where they can copy and send the data to their retrospective analysis device, analyzer console, or a third-party purpose-built device.” Thompson told ESJ that overhead is extremely small and will not disrupt application performance.

Users can access a full and integrated view of virtual traffic as it crosses physical networks, moves between virtual machine hosts, or travels between virtual machines running on the same host. All virtual traffic and communications tracked inside the VM host can be sent to the company’s GigaStor appliance, which can be used to analyze historical activity to determine problems and propose solutions. This “enterprise-wide views of virtualized resources and performance within virtualized and physical environments” is what sets the product apart, the company says.

Thompson remarked in a prepared statement, "Companies often migrate to virtualization to cut costs, but find the move opens up gaps of coverage as their traditional monitoring tools can't address these new environments. Customers can rely on the Observer platform to provide visibility to all points on their network, regardless of whether it's at the edge or core or within physical and virtualized environments. Using our performance management platform, engineers can pinpoint performance problems from high-level reports and immediately perform root-cause analysis."

The company says that current maintenance customers are eligible for the product at no charge. More information is available at www.networkinstruments.com.