Improving wide area network (WAN) performance requires centralized views from branch office network monitoring...
tools. See part 1 of this article to learn about NetScout's branch office network monitoring probe in Cisco's ISR, or continue reading below to learn about other branch office network monitoring options.
Some enterprises have tried to gain visibility into branch office network performance by turning on Netflow, Sflow or some other kind of flow-based monitoring technology in the branch routers and feeding that information into a centralized monitoring tool, according to Jim Frey, research director at Enterprise Management Associates.
"This gives you some information, but it doesn't give you the same depth of information as a [packet-based view]," he said.
Turning on something like Netflow in a branch office network router can also affect router performance, according to NetScout vice president of marketing Steve Shalita.
Application performance monitoring vendor Opnet has partnered with WAN optimization vendor Riverbed Technology in a fashion similar to the NetScout and Cisco collaboration. Opnet's ACE Live performance monitoring product can run locally on the Riverbed Services Platform (RSP), a VMware-powered virtual partition on Riverbed's Steelhead WAN optimization appliance that serves as an alternative to the ISR service module. With the Opnet product running locally on the Steelhead, the virtual instance of ACE Liveit can then send detailed WAN performance information to a centralized monitoring console at the corporate headquarters.
The value of a centralized view of branch office network monitoring
Frey said a seamless and centralized view of how the network is performing, from the data center to the branch office network user, is critical for a number of reasons. A WAN manager could solve branch office network problems on the fly by using some local, disconnected network management tool or by logging into the router remotely and checking the logs. He could also send someone to the location with a packet sniffer to figure out the problem. But these approaches are time-consuming and don't fit into a seamless and integrated network management approach.
"It really comes down to how quickly you can figure things out when they break and what you are willing to risk in terms of mean time to repair. Also, a lot of the focus is moving away from simple break-fix and more toward predictable, reliable performance, which is much more proactive and requires you to do sustained monitoring and trending analysis," Frey said. "That can only happen when you're bringing data in from the same sort of ongoing performance measurement points back into the main system so you can apply some sort of trending analysis. If you only have a disconnected network monitoring element that you can't pull up in the context of the rest of the system, you're not going to have that ability to analyze performance trends."
Centralized branch office network monitoring with good trending analysis capabilities is also critical because break-fix issues aren't so much the focus for enterprises anymore, Frey said.
"The big problem out there isn't, 'Is my application available?' It's, 'Why is it slow?'" he said. "There are a lot of reasons why that could be the case."
Frey said enterprises struggle with determining whether slow application performance is unique to one branch or part of a bigger problem. With local branch office network monitoring probes embedded in platforms like the ISR or Riverbed's Steelhead, enterprises can pinpoint the problem much more quickly.
Let us know what you think about the story; email: Shamus McGillicuddy, News Editor