NetworkVantage (NV) is a performance monitoring tool whose data collection and reporting is based not just on low-level link throughput but also on the various applications used within the enterprise. The reasoning is sound: although network managers care about the performance and utilisation of the network itself, the ability to put network usage in the context of the application traffic that sits on top of the network allows a greater understanding of the cause of any performance issues that arise.

NV works by analysing network traffic from around the corporate network. It collects data via agent programs, or probes, that sit on computers on the various network segments; these probes are managed from the first of the front-end applications that are copied to the host machine by the installer: the NV Console.

The Console program is a fairly basic two-pane application that allows you to tell the system where the various probes are. You can implement a hierarchy if you have a lot of probes, nominating some machines as probe managers and connecting the NV Console to these machines instead. As well as collecting data from the various probes, the NV Console can send patches to the remote device, for instance, if a new version of the software becomes available.

As well as allowing probe configuration, the NV Console also lets you manage the list of applications that the system knows about. The system recognises a decent number of applications as standard but there are some omissions (it didn't seem to recognise Windows Terminal Services traffic in our test, for instance). So you can "Define application" and tell the system how to identify stuff based on addresses, ports, RPC ID, SOAP action values or even IPX sockets.

Once configured, the system will collect data from the various agents, ready for you to start viewing. This is where the second GUI application, the Interactive Viewer, comes in. This allows you to dig into the individual collection files that the main package produces from its agents, so you can find out what's going on. You can look at the data by application type, device, link or computer (layer address or MAC address) and within each section you get a sensible set of sub-categories, so if you're drilling into the data based on computer address, for example, you can choose from AppleTalk, IPX, IP and so on.

Most of the data visualisation is graphical, and it's a true drill-down system in that if a section of a graph relates to a machine, you can double-click on that to be taken to the data specific to that machine. Rather like a Web browser, there's also a back button in case you get a bit lost moving around the graphs. As well as viewing the data interactively, you can define and produce multi-page HTML-based reports that include as many screens from the Interactive Viewer as you wish to tick in the report creation wizard.

The only slightly clunky bit of NetworkVantage is the way that the Interactive Viewer works with individual collection files - when you run the program, you open the collection file that corresponds with the dataset you wish to view. Given that NetworkVantage stores much of its data in a database, it would be more elegant for the viewer to interface to this database and produce a list of available datasets, instead of requiring the user to navigate down the Windows directory structure to find "whatever.vdb".

On the whole, though, NetworkVantage is a useful package that provides more than your average performance monitoring tool through its ability to list data not just by machine or link but by application.


Remember that with a back-end SQL Server database running alongside the package you'll want to run this on a half-decent server to get reasonable performance.