Network Physics: Looking Deeper Into The Network

Even though the company name sounds more like a college class than a networking hardware vendor, Network Physics is indeed the latter, selling the NP-2000 appliance, one of several entrants in the growing field of network traffic diagnostics.

According to the company, its "ground-breaking physics research" is both the basis for its name, as well as its product's ability to detect and diagnose IP network traffic flow patterns in real time without adding to bandwidth constraints. Networking Pipeline recently talked with David Jones, Network Physics' CEO and president, about the metrics measurement market, and where Network Physics fits.

Networking Pipeline: How do network managers compute the return on investment from something like the NP-2000? Is it just from being able to replace hardware, like network probes?

David Jones: There is very definitely a strong ROI in the reduction in probes. But generally, the capital equipment cost in a network is trivial. It's about 20 percent of your budget, versus 80 percent of operating costs. On the operating costs our product offers a very definite strong return.

It's really more about control: How do place a dollar value on the resolution of an issue that's affecting your entire supply chain? What is the difference between resolving that in an hour, a day, or a week? When you have a business-critical network, it can't be down. If you have a problem, you have to be able to put your finger on it.So I don't know if this is classic ROI, but clearly, there's ROI. And what we're seeing from customers is that motivation today is less ROI, and more 'I gotta keep my apps up and running.' For the CIO, the next call he gets from the CEO is not to squeeze another dollar out of costs. Instead, it's 'how come my Exchange is so slow on Monday morning.'

Networking Pipeline: Do users want a device that just gives them metrics, or do they want something that also lets them take action?

Jones: In fact it's just the opposite, particularly when you are dealing with management of routers. What users want out of a device is, 'tell me where there's a problem, tell me why there's a problem, and let me evaluate it -- then I'll do something.' There are lots of examples from the early route-optimization [product] days where re-routes would happen automatically -- with very severe and unintended consequences.

Clearly the path we're traveling is to automate [the process]. But the way we'll get there is a device that isolates what's happening for you, and then is suggestive as to what you might do about it. But the first step will be the 'do it' button.

Networking Pipeline: To keep administrators in the loop?Jones: Yes -- remember, we are dealing here with very complex environments, so it's hard to automate processes. The other place this is going to come from is not on this side of [highway] 237. [editor's note: Network Physics' headquarters in Mountain View, Calif., is across Hwy. 237 from Cisco Systems' nearby campus in San Jose.]

Networking Pipeline: So this product [the NP-2000] could be a blade, inside a Cisco switch or router.

Jones: Absolutely.

Networking Pipeline: Is that part of the plan? Have you had those conversations yet?

Jones: We have had those conversations here [at Network Physics]. But I'm not sure the market's right yet [for an integrated product].0