The odds are if you're running a corporate network, you're wasting money somewhere.
But Carrie Higbie said that doesn't have to be the case. And when Higbie talks, people listen. She's been in the industry for more than 25 years, participates with various consortiums for standards acceptance, has a massive background in all aspects of networking development and is an expert on SearchNetworking.com, SearchEnterpriseVoice.com and SearchDataCenter.com.
Along with her credentials, Higbie knows how to ensure a network is cost effective and running at its full potential.
Recently, Higbie put together "Cut Costs with Network Performance Management," a presentation to help enterprises measure their networks' performance while thwarting unnecessary spending. In an interview with SearchNetworking.com, Higbie touched upon some of her main points.
Higbie, a global network applications market manager to support the end-user and electronics communities for the Siemon Co., said the first thing a company should do is study who and what is using the network and how they are using it.
"You have to take studies into account when you're doing a utilization study," she said. "The idea is it needs to be a study over time. If you don't plan that a peak is going to happen all the time, something is going to suffer."
Other areas performance management tests can check include Web page, LAN, WAN, VPN, application and cabling performance. Some performance problems may be unknown, she said, and can be fixed with a little house cleaning.
Higbie said performance can take a hit with poor planning. Certain roots that lead to bad network performance include poorly installed infrastructure, poorly written code, server side services, WAN or provider side services, undesirable traffic, poor virtual LAN management, addressing errors, poorly maintained cabling, too many or unnecessary protocols or an oversized address pool.
To save money, companies must also consider moving some ports around, moving devices that are bandwidth hogs to devices with spare bandwidth. It also helps to determine if the infrastructure is using unnecessary bandwidth and if devices are bullies.
"It may not be that you need new purchases at all," she cautioned. "Don't buy technology you don't need. Don't buy new technology for the sake of buying new technology."
Higbie cited a recent study that suggested 20% of all IT purchases were for products that, for some reason or another, didn't work. To ensure that doesn't happen, Higbie suggested companies develop a five- to 10-year technology plan to determine what they need now and what they'll need in the future, instead of just using a quick Band-Aid to fix any problem that arises.
A technology plan, Higbie said, should look at technology progressions and the company's expectations, and account for trends and growth. Higbie said discussions about technology plans often fail to include every department, which does not use all the resources and knowledge available. Plus, she said, some departments may have tools another department can use, which avoids added costs.
"If you're going to do a study for planning purposes, you need to have different departments involved," she said. "The biggest recommendation is have someone from every department there. Everybody's got different budgets. Everybody has to get together to decide what's best for the company."
"When things just stop working, it's a reactive approach," she said. "You can Band-Aid a lot of things, but that doesn't heal the wound. Proper planning and proper auditing of what's in place can save a fortune down the road. Downtime is ridiculously expensive."
Overall, Higbie said, companies must look at what they have and how it is being used. From there, they should evaluate how it can be used better. Then, and only then, should an enterprise decide to make expensive changes.
"To be most effective," she said, "you have to step back before you look forward."