Sneak Preview: Network Physics' NetSensory Enterprise Architecture 5.0 Provides Insights Into the Network

IT networks run 24/7, making it difficult to get at the root of almost any problem. Network Physics has added NetSensory Insights, best-practice templates that offer workflow views to help identify trouble spots, to its NetSensory Enterprise Architecture 5.0, which is touted as more than a network-management platform because it monitors the entire application-delivery infrastructure.

NetSensory Enterprise Architecture is based on a distributed operating system that runs on both the vendor's NP-Director global-management appliance and its NP-2000 data collectors. The management GUI is in a Java application that downloads automatically to your admin desktop.

I tested the NetSensory architecture in our Syracuse University and Green Bay, Wis., Real-World Labs®, using NP-Director and two NP-2000s. The appliances were all delivered on standard 1U Intel hardware.

Before starting, I ran an upgrade to NP-Director and both NP-2000s--a time-consuming manual process. I wish NP-Director could have been scheduled to push the files out automatically.

Application Scanning

As soon as I plugged the appliances into the network, the NetSensory architecture scanned for applications. It identified well-known ports, searched for server applications requiring a range of ports, and sought out user-defined applications that use ports, servers and URLs in combination.

Once the network is defined and the applications are identified, the NP-2000s collect network- , server- and application-performance data much like an RMON probe does. This method requires no agent and uses little overhead.

The NP-Director server aggregates the data and presents it on the Java interface. Trouble is, with so many drill-down options, you'd have to be a network engineer to know where to begin. For this reason, Network Physics has added five predefined, customizable NetSensory Insights views--Audit, Baseline, Optimize, Troubleshoot and Security--to organize the performance data for you.

Added Insights

Audit Insight displayed a performance view of applications, servers and clients. The discovery parsed packets on the wire and found all traffic going to and from the Syracuse labs, as well as critical servers and a mess of clients.

Next, I examined Baseline Insight, which breaks out trends monthly, weekly or daily, depending on how you configure it. It showed me performance statistics for nodes with the highest bandwidth usage, the most latency and the greatest packet losses.

Optimize Insight displays chatty applications. It lets you compare their performance before and after a change by selecting key statistics, such as payload and packet rates.

Troubleshoot Insight identifies bandwidth hogs and slow servers. Security Insight uses traffic patterns to look for worms.

Sharing With Others

This version of NetSensory has a read-only option for Insights, letting service desk and management folks log in to check out designated performance data. All NetSensory screens can be e-mailed outside the company.

Overall, Insights improve NetSensory by making it easier to combine and compare information. But for network diagnostics, you'll still need someone smart running the show.

Bruce Boardman, executive editor of Network Computing, tests and writes about network and systems management. Write to him at [email protected].