A look into Network Physics’ crystal ball

While still a relative newcomer to the performance management marketplace, Network Physics touches one of my personal hot buttons: leveraging insights across management disciplines to enable multiple types of decision-making.

In an appliance design, Network Physics offers fault-, performance- and usage-based analysis for accounting, security, and service management – across private and public networks, with insights into networks, networked devices, applications and service/business groups.

The NP-2000 appliance provides performance metrics such as response time, hop-based latency, and packet loss; utilization information for specific application flows; BGP and ISP route mapping and troubleshooting (route analytics); and packet capture. Its design – originally focused on Internet and WAN challenges, such as “invisible” VPN transmissions – is fairly fluid across VoIP, wireless, business applications (for example, ERP, Web services and so forth), VPNs, and MPLS-based services. Network Physics likes to characterize its product as “end-to-end,” Layer 4-directed “Business Network Integration.”

One of Network Physics’ advantages – as long as it continues to harvest this advantage with more advanced reporting and GUI capture – is an integrated database that can allow for flexible, mix-and-match visualization across these different data types. The NP-2000 can organize data according to business group, application group, IP conversations and other types of groups. This means a relatively fluid path for growth in supporting network, application, security and business professionals across a range of decision-making, from planning through operations. Looking at this data as an integrated tapestry, however, sometimes causes viewers to pause – as if they were looking at a crystal ball for the first time and wondering just how to use it.

Enterprise Management Associates’ experience with Network Physics customers has been positive. Most remain in discovery mode, which is normal for an early-phase, enterprise management product with broad applications. In one environment with a global network, filters were developed in conjunction with Network Physics for virus protection, based on the Network Physics appliance’s capabilities for determining anomalous application patterns in conjunction with normal behaviors. Troubleshooting and service support for asymmetrical applications (including streaming media) across widely dispersed WAN users was also seen as a consistent plus. And in various cases, the Network Physics appliance was used in conjunction with other enterprise management products – such as those from Opnet, Micromuse, InfoVista and NetIQ – to provide a more automated foundation for production-level control of a complex, distributed infrastructure.

One of the requirements for Network Physics going forward, and the company is aware of it, is the capability to integrate distributed appliances across a larger network, for single, integrated views. Currently, the NP-2000 is placed at the edge of a data center targeted at monitoring WAN and local campus interaction from a single, strategic point. As it is agentless, installation is relatively easy. The appliance collects and assimilates data, and administrators can deploy policies for prioritizing what types of views are preferred. Flexibility, breadth and a more defined focus in how data are presented and organized are clearly other areas of growth for Network Physics – areas that, over time, should help to clarify the full dimensions of its “crystal ball.”

Pricing for the NP-2000 depends on the bandwidth managed; the average initial deployment consists of two or three units and runs between $100,000 and $200,000.