News Stay informed about the latest enterprise technology news and product updates.

Traffic shaping helps cure network congestion woes

Traffic shaping and caching technologies, which prioritize data transmissions based on type and user, can help ease -- and dramatically lower -- network congestion. The typical enterprise network is becoming a more and more crowded place: Every day, a new video sharing site, communications platform or Web 2.0 wonder pops up and claims its share of bandwidth. Find out how one library system has used traffic shaping technologies to help solve its bandwidth challenges.

The typical enterprise network is becoming a more and more crowded place. Every day, a new video sharing site, communications platform or Web 2.0 wonder pops up and claims its share of bandwidth. Traffic shaping, or prioritizing data transmissions based on type and user, is now a critical tool for many network administrators.

Those with congested networks shouldn't feel bad: Industry analyst group Nemertes Research recently published its findings that the entire Internet may soon hit some bottlenecks, so at least congested companies are not alone. In fact, the two problems may have quite a lot in common.

"The edge of the Internet acts like a LAN," said Mike Jude, senior analyst at Nemertes. "It turns out there is a fair amount of concern about the degree to which extraneous traffic is choking corporate networks."

Jude identified a few key sectors -- universities, enterprises and government entities -- that were among the first to feel the crunch, but it is a concern across the board for administrators, he said.

Often, the best approach is to study traffic patterns and then suggest policies, he said. Are large emails draining bandwidth? Is YouTube slowing your network connection? This kind of data is invaluable when planning your attack strategy.

More on network latency and bandwidth management
Guide to bandwidth measurement and management

Bandwidth: How much is enough?

What is network latency?

How can I tell when packet loss is affecting application behavior?

Assessing WAN connectivity, identifying latency for centralized application access

Help prevent latency problems by designing your WAN effectively

Sean Hill is one network administrator facing congestion with some tough constraints: He is the information technology manager for the Henderson, Nevada, library system, which includes three branches and 250 PCs, about half of which are used by the public for everything from homework research to running a home business. "When I talk to colleagues, they are all not letting people do things, but we're all about letting people do things," Hill said. This includes social networking, viewing video clips, and all the other high-bandwidth activities that can send a chill down a network administrator's spine.

"Because we're a library, I'm not stopping people from going to those places," he said. "They're allowed to go wherever the Internet takes them."

On top of its inclusive use policy, the library is moving to pure VoIP telephony and has a videoconferencing room, neither of which can work with high latency. The performance of the latter was threatened a few years ago as the network experienced growing pains.

"There was a while there when I had to say, 'We can't afford to do video calls ... in the afternoon,'" Hill said. That has changed since he started using a combination of packet-shaping and compression tools to ease the burden on his WAN.

About five years ago, Hill deployed Packeteer's iShaper platform to help dedicate bandwidth where it was needed while reducing redundant data transmissions from emails and file shares. He says that both have given his network a huge boost.

File compression for the library's numerous shared documents saves about 50% of bandwidth, Hill said, greatly improving responsiveness for employees.

But perhaps iShaper's more critical feature is its ability to detect and prioritize traffic types over a WAN.

"If somebody is streaming movies on YouTube, that's fine with me," Hill said. "I don't want that streaming affecting other people."

iStreamer allows Hill to prioritize all traffic on a scale of 1 to 7. VoIP and videoconferencing are prioritized 2 or 3, Web traffic might be prioritized a 4, and then specific sites like YouTube or Internet radio stations are put on a "high bandwidth" list that Hill made and prioritized 6. Those sites will still go through, but at a higher latency to allow for other services to run uninterrupted.

"On some days that might mean it's not usable," he said. "On other days, the packet shaper lets them have almost the entire Internet link if we want it."

In addition to the by-protocol limits, each individual user can then be capped for more fine-grained control.

Hill said he once needed a 10 Mbps connection, but because traffic shaping has lowered the excess bandwidth he needs to keep handy, he now feels comfortable with a 5 Mbps optical Ethernet connection.

"You're paying for the device pretty quickly when you start adding up five connections at $500 a month," he said.

Jude said that one of the keys to making an implementation such as the Henderson Library deployment work successfully was adapting it to the network's needs.

"One of the things that you're not really sure of is if your employees are using YouTube to talk to each other … there may be some real business legitimacy to that [which] management might not immediately see," he said. So banning YouTube outright could actually hurt efficiency.

In Hill's case, regular monitoring and prioritization helped balance the need to allow a free flow of information with the need to make sure that information doesn't overflow the network's capacity.

Dig Deeper on WAN optimization and performance

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.