Paul Langway, Network Instruments premier dealer with reseller PVP Sales, agrees. "The integration of OI into their aggregated high level reporting provides the user the ability to understand at a glance the root cause of the problem that often looks like a network issue, but ends up being an issue in the application or the platform the application is running on--such as processor utilization, available memory, load balancing issues," he says. According to Langway, there are many solutions that identify device issues or application performance. But most don't aggregate the information in a single tool, providing the user "with a top-down approach to drill down to the specific area from a single report."
Observer Infrastructure also is designed to help organizations monitor performance of VMware's virtual machines (VMs) by discovering and monitoring VMs and virtual hosts to poll their CPU and memory and track utilization and resource consumption metrics. "Companies want to be able to drill into that opaque environment," says Network Instruments' Reinboldt.
Managing virtual and cloud-computing environments remains a challenge for many organizations, says EMA's Frey. In particular, it can be difficult to monitor virtual traffic on the wire, between hosts and even between virtual machines within a single host. Network Instruments' offers Virtual TAP functionality (announced in March 2009), which is designed to direct all traffic to a dedicated virtual NIC then mirrored on one of the host's physical NICs so it can be analyzed and monitored.
While Frey says Network Instrument's Virtual TAP is a good approach, there are still hurdles. "Most shops aren't willing to dedicate VMs to these types of probing. VMs are often under the control of systems administrators and they don't care about or know about or want probes on the VMs," he says. Available now, the Observer Reporting Server appliance costs $25,000. The Observer Infrastructure appliance is priced at $17,500, and to additional appliances or software range in price depending on the number of devices managed.
