Meru Networks today introduced version 2.0 of its E(z)RF Network Manager. The update represents a major step forward for the solution, adding continuous event recording and proactive alerts via a knowledge-based inference engine, as well as a mobile interface formatted specifically for the Apple iPhone and iPod touch.
According to Kamal Anand, Meru’s senior vice president of marketing and corporate strategy, the key strength of the new version lies in the fact that it continuously captures data for every device on the network, and then checks that data against a knowledgebase of over 100 potential client and infrastructure issues, generating automatic alerts as needed.
And that’s a significant shift from the usual network management approach. “We are continuously capturing over-the-air events and client state transitions,” Anand says. “It’s not a periodic poll, it’s not aggregate statistics… what we are able to provide is all the state transitions that every client went through in the network—and we are also monitoring the infrastructure itself, the access points and the controllers… and all this information is then put centrally in a database.”
The resulting data can then be viewed in as much or as little detail as needed, even down to the individual client—and can also be ‘rewound’ to a previous state to diagnose past problems. “As an IT manager, when I get a trouble ticket at 10am from somebody who reported a problem at 4pm, I can with two clicks enter the MAC address of the user and roll the network back to 4pm last evening—and I can see exactly what the client was doing in terms of the state transitions, and I can see what the rest of the network was like… and I can view all the events that are going on, which might reveal the trouble that was happening at that point,” Anand says.
The idea is to save an admin from having to troubleshoot by going through every possible cause of the problem. If a user can’t get online, for a example, “you can rewind and recreate the network that existed in the past, and you can very quickly see that the user was trying to associate, but the RADIUS server was not responding… and it might take you five minutes to figure this out,” Anand says. “So the end result is, you can do this much faster—and the end user can be up and running much faster.”
At the same time, all data is also automatically checked by the inference engine. “We’ve built a knowledgebase of typical problems or patterns that might lead to a particular problem—and so this is continuously going on in the background: the inference engine is parsing all this data… and if it notices an anomaly or a particular pattern, it can raise a notification to the IT manager,” Anand says.
The result, according to Anand, is a more proactive approach to network management than simply waiting for users to call in with issues. “A lot of time, a user might get a disconnect and say, ‘Okay, I’ll wait for half an hour and come back’… and firstly, that’s downtime, and secondly, it might be a symptom of other problems,” he says. “So being proactive, we believe, allows the network to stay up much more for all the business critical applications.”
And the iPhone interface, Anand says, provides mobile access to all key aspects of the system. “All the critical events, alarms, and dashboards are available—they’ve been formatted specifically for the iPhone,” he says. “And it’s a browser-based application, so you don’t have to load anything on the iPhone itself… and we can continue to enhance the application without uploading to the iPhone in any way.”
Anand says the new version of E(z)RF Network Manager can be of use to a wide range of employees performing a wide variety of functions, “from basic network administration, health checks, long term planning around capacity… to doing helpdesk troubleshooting with very detailed packet captures, rewinding and recreating the network, proactive diagnostics, all the way to senior executives looking at snapshots of the data on their iPhones.”
Looking forward, Anand says Meru will continue to enhance the inference engine with more patterns and events, and will also work to expand the mobile solution to other smartphones and mobile devices. “We can also start providing APIs to the database for third-party scripts… so we can have other parties mine this data for other uses that we may not have thought about,” he says.
E(z)RF Network Manager 2.0, which can scale to support up to 25,000 access points, is available now. Software supporting 50 APs is priced at $4,995, and requires a Meru SA1000 Service Appliance, priced at $6,995. The smartphone solution, E(z)RF OnTheGo, is priced at $499 per user license, though it can be downloaded for free during an initial promotion (until September 2009) by any Meru customer with an E(z)RF Network Manager 2.0 service contract.
Jeff Goldman is a veteran technology journalist based in Southern California. He is a frequent contributor to Wi-Fi Planet and other Internet.com publications.