Network management tools are supposed to help administrators take care of their companies’ infrastructure by flashing hazard signals. Protocol analyzers and packet sniffers flash warning lights when things go awry. But thanks to increasingly complex network architectures and staff shortages IT departments, those colour-coded light can easily become more of a hindrance than a help.
The red and yellow indicators are supposed to show snags in the network and help managers decide how to repair traffic flow trouble. But for all of the high-tech whiz-bang of the software, managers sometimes ignore these tips.
“It’s a sad state of affairs,” notes Glenn O’Donnell, product manager with the Meta Group Inc., a Palmerston, Penn.-based research firm. “A lot of the tools being used are either not capable or not configured to do event correlation – the process of eliminating all of the false alarms and getting to the route cause.”
But today’s diagnostic technology is only one part of the problem, he said. Low staff levels and poor policies also hinder network management.
Those alarms offer little respite when there’s no one available to attend to them, said Brad Masterson, spokesman for Mississauga, Ont.-based Fluke Electronics Canada.
“This issue has been around for quite some time,” he said. Hiring trouble began when IT experts were hard to come by, back when there were too many jobs for not enough technologists. “Now there’s a lack of IT staff because [IT departments] just can’t hire,” since companies are trying to keep spending down.
With few staff members to spare, network managers have trouble baselining (finding out what represents normal levels of activity) the infrastructure and updating the management software with current threshold information.
“You have to baseline the network…(to) get an understanding of what’s on your network,” Masterson said. “If you don’t know what’s normal today, when you go to do that baseline later you won’t know what’s really changed.”
And without baseline information, the network management software doesn’t know what “normal” entails. The situation renders the application useless and makes network management that much more difficult.
The solution lies in a cohesive view of the network, said O’Donnell.
“For example,” he said, “incident management should be centralized across all of the technology silos.” The analyst advocates a grand network operations centre, an overarching command point that offers a complete view of network architecture.
He added that it isn’t difficult to convince the purse string grips to let loose the cash, provided they’re given reason enough to invest in network management.
“You have to have some solid return on investment analysis done on it. If you can prove that it will, indeed, save money or somehow increase revenue for the company, it comes down to the business side. They’re going to spend the money.”
But others say the solution to useless alarms need not be so intricate.
Paul Thiessen is the network administrator at the Network Management Centre, an IT outsourcing company in Cranbrook, B.C. He figures the answer is simple: recognize your company’s limitations when it comes to staffing levels, combine that with the software’s limitations – and deal with it.
“Some software does have the capability to be modified or configured to the full extent, but others don’t. They can’t all be all-encompassing. It would be nice if you had a tool that you could configure exactly to your needs in the organization, but you have to look for a best of breed or a blend of tools.”
If all else fails, “take everything with a grain of salt,” Thiessen suggested, “and use your own judgement as to the extent to which you follow through on the information you’re receiving from the technology.”