Bandwidth Management Coming To You

Cox Communications recently announced a new bandwidth management program, while Google and partners are releasing a tool to detect throttling. The traffic battles are heating up, but the deck is stacked against users since we use the pipes, not manage them. Even so, Cox's plan seems responsible and, if done right, can balance competing network demands.

Mike Fratto, Former Network Computing Editor

January 29, 2009

3 Min Read

Cox Communications recently announced a new bandwidth management program, while Google and partners are releasing a tool to detect throttling. The traffic battles are heating up, but the deck is stacked against users since we use the pipes, not manage them. Even so, Cox's plan seems responsible and, if done right, can balance competing network demands.Face it, there isn't an infinite amount of network bandwidth available. As more traffic streams are aggregated together, you, and everyone else in your neighborhood, town, and region, are eventually sharing the same resource. Throwing stones at Cox for not planning for unlimited bandwidth is futile. It should have the capacity to support the traffic claims made in marketing literature, but that the solution is for an ISP, any ISP, to accurately specify bandwidth, not maximums, or worse, unlimited bandwidth, is ludicrous. But marketing is marketing and consumers want more for less.

Other than continually adding bandwidth, the next best thing is to manage the bandwidth you have. Congestion has a greater affect on real-time traffic like voice, video, and gaming, than it has on other traffic like HTTP, file transfers, and e-mail. Who cares if your file transfer has a latency of 150 ms and a variance, jitter, of 30%? It won't have much affect on transfer rates but it will ruin that call to grandma or on-line gaming. I'd hate to be in a firefight just to get shot because of network lag. If done properly, congestion management kicks in when the network becomes congested and priortizes some traffic, like time-sensitive traffic over bulk transfers. If there is no congestion, then bandwidth doesn't have to be managed. Traffic shaping is a good thing and a commonly used method to better utilize a scarce resource.

Looking at Cox's FAQ, it lists the types of traffic it considers time-sensitive, such as Web surfing, IM, and real-time media, and non-time sensitive such as file transfers and P2P activity. It also states "any traffic that is not specifically classified will be treated as time-sensitive." It appears that Cox's plan is to de-prioritize traffic that it has identified as non-time sensitive and let the remainder duke it out.

It's an interesting stance. It could have done the opposite and identified the traffic that would have to be prioritized, but I think de-prioritizing bulk traffic is a simpler and cleaner solution. File transfers have a distinct traffic profile and can be easily and reliably identified using a variety of methods. But new applications and protocols are popping up on the Internet, so chasing the next new thing would be time consuming for it and potentially disruptive to you.

The key for Cox Communications customers, and customers of other ISPs, is to verify the ISPs are doing what they say. Google and the Measurement Lab have released some tools that are supposed to test various types of traffic to detect throttling. I haven't tried them yet. I suspect that they are slashdotted even as we speak. But the tools will test network speed, whether P2P is being throttled or blocked, and other diagnostics.

Take the results of speed tests with a grain of salt. Internet speed tests look cool with moving gauges and the like, but don't really tell you the full story because there are a number of factors that affect performance such as path, protocol, and all the infrastructure end-to-end. Your biggest congestion point may be the air around your 802.11 wireless access point.

Congestion management is coming. Get over it. What is critical is that congestion management is done properly and responsibly.

Read more about:

20092009

About the Author(s)

Mike Fratto

Former Network Computing Editor

Mike Fratto is a principal analyst at Current Analysis, covering the Enterprise Networking and Data Center Technology markets. Prior to that, Mike was with UBM Tech for 15 years, and served as editor of Network Computing. He was also lead analyst for InformationWeek Analytics and executive editor for Secure Enterprise. He has spoken at several conferences including Interop, MISTI, the Internet Security Conference, as well as to local groups. He served as the chair for Interop's datacenter and storage tracks. He also teaches a network security graduate course at Syracuse University. Prior to Network Computing, Mike was an independent consultant.

Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.

You May Also Like


More Insights