Burstable bandwidth

01 January 2011 |


Burstable bandwidth is a means by which the bandwidth offered by a service provider can be boosted to address peak periods of usage.

By increasing the capacity available in a burst to address that usage, customers don’t need to sign up for a continuous and more expensive higher level of service. In addition, providers don’t need to scale their networks to support vastly higher peak usage rates that will not be used for much of the time.

Burstable bandwidth describes a connection in which the customer’s available bandwidth can be increased for short periods of time on an as-needed basis. Service providers that offer burstable bandwidth with a package will let customers make use of some or potentially all of the redundant bandwidth available in their network to help them cope with the traffic spikes and other forms of high bandwidth demand.

What are the advantages of burstable bandwidth for the user?


By purchasing a plan that allows for burstable bandwidth, users can prevent their sites slowing down during peak traffic periods. Burstable bandwidth means they can use more bandwidth than their hosting plan normally allows them to. Burstable T1 lines also offer substantial cost savings to businesses that require full T1 bandwidth but only a few times a month. Burstable T1 connections are capable of full T1 bandwidth when required but reduce available bandwidth when full capacity is not needed.

Bursts of additional capacity can only be provided where there is redundant capacity available so the ability to burst may not always be there.

In addition, burstable bandwidth should really only be used as a short-term solution because it is much more cost-effective for users that consistently exceed the limitations of their bandwidth package to upgrade to a higher level.

By far the greatest downside to burstable bandwidth is its cost. While it makes an excellent proposition for businesses that require the utmost in scalability, for others, the threat and unpredictability of higher costs makes it unappealing.

How is burstable bandwidth charged for?


The usual method is using burstable billing – which burstable bandwidth is also sometimes called.

Typically, burstable billing is based on 90% or 95% of the utilisation of a network connection. By ignoring the top 5%, the needed capacity of a link is reasonably uncovered. The 95th percentile allows a customer to have a short – typically less than 36 hours per month – burst in traffic without charges for exceeding their agreed usage. This approach means that 95% of the time usage is below that agreed amount and 5% of the time traffic may be bursting above that agreed rate. Some providers offer billing on the 90th percentile.

What are the costs involved in a burstable bandwidth service?


The nature of burstable bandwidth means that it is often hard to pin down in advance the costs of a service.
 
For example, a user that has purchased 3.0Mb of committed bandwidth could still be able to burst up to 10Mb or more depending upon their original sales agreement.

If the company bursts excessively beyond its committed bandwidth, it will incur an overage charge, or fee for the extra bandwidth it has used. These overage charges are calculated by using switch readings that typically take place every five minutes. As a consequence, 8,000 to 9,000 readings are gathered by the service provider each month and stored in a database. The top 5% of these readings are discarded to create the 95th percentile.

An example of how burstable bandwidth charges are calculated would involve a customer purchasing a 2Mb per month guaranteed bandwidth service for $350. Then at the end of the month, if the customer’s total use is 3.7Mb based on the 95th percentile then the overage is 1.7Mbs. An overage fee rate could be $250 per Mb or greater and that, multiplied by 1.7 to represent the 1.7Mb overage, provides the overage fee of $425, giving a total cost to the business of $775. That provides a clear figure from which the customer can therefore determine whether to pay a higher fee for a greater amount of guaranteed bandwidth or whether it would make sense for them to continue to rely on burstable bandwidth to address irregular peaks of usage.