Nag, nag, nag: Monitor those WAN apps, already!

Over the years, we’ve often lectured you about the importance of monitoring the utilization of your WAN access bandwidth and the application traffic traversing this network segment. The access network is where capacity is usually most constrained and, thus, where bottlenecks are most likely to occur.

It’s time to get on our soapbox again.

In a recent survey of about 190 Network World subscribers, co-sponsored by traffic-shaping appliance maker Packeteer, well over three-quarters of the companies with revenue above $1 billion said application performance degradation had at least moderately impacted employee productivity, team productivity and customer service. They rated the WAN as the No. 1 cause of the degradation, and more than half the companies of this size said the degradation had worsened over the past 12 months.

However, more than 75% of respondents indicated they were pretty much in the dark as to what applications are even running on their networks, stating that their knowledge ranged from “know some” to “do not know.” Put another way, just under one-fourth said they “know precisely” what applications are running and how much bandwidth each is consuming.

It’s pretty difficult to ensure application service levels without knowing what other applications are out there impacting your most critical ones.

When we began writing this newsletter five years ago, we advocated capabilities in then-emerging intelligent DSU/CSUs, which monitor your WAN traffic and generate reports on how well your applications are performing. Only when armed with the facts can you take appropriate action.

Today, in addition to the smart DSU/CSU options, there are sophisticated monitoring and traffic-management capabilities from several types of companies. These include appliance makers such as Packeteer, Allot Communications, Expand Networks, Peribit Networks and Sitara Networks, as well as software-based alternatives in WAN access routers from companies like Cisco. These vendors have combined granular monitoring of traffic by user, interface, protocol, application and port (and applications that can hop from port to port), so you can view application behavior based on these variables.

Then you can prioritize traffic, nail up a minimum amount of bandwidth to certain applications (such as voice over IP), rate-limit bandwidth-hogging traffic, disallow certain traffic and, in some cases, compress traffic to squeeze more out of existing bandwidth.

We’ll take a closer look at these issues next time.