Bandwidth is like the soul -- everyone thinks they know what you mean when you discuss it but they are probably thinking about something entirely different. And that's because, like the soul, it does not transparently exist in any of the simple ways that we tend to believe.
Why begin this piece with bad metaphysics?
For a start, the word "bandwidth" is not even semantically sound -- it's a vague label improperly attached to a range of meanings (see 'Bandwidth is a misnomer' for more insight). Implicitly, it most typically refers to the maximum rate at which data can be transferred, either at a single interface or across an end-to-end path. But even this straight-forward idea of "maximum data transfer rate" can be endlessly qualified. Anything from packet size, to cross-traffic, to corruptions can limit the effective transfer rate.
Somehow though, it remains the central characterization for most networks. As if size matters. And it does. Sort of.
Consider that the raison d'etre of any network is NOT to transfer bits -- it is to enable applications and end-users. It is easy to imagine a network that has lots of "bandwidth" but fails to support the applications using it. VoIP, a low-demand application requiring very little bandwidth, has made that abundantly obvious.
So the functional capacity to transfer data is not the sole metric for network performance. Certainly the required transfer capacity must be present -- but other aspects of performance like loss, latency and jitter may be far more important. And this is why an increasingly sophisticated relationship between network behaviors and application performance is beginning to emerge (see 'Making the triple play').
But still the rate of data transfer is critically important to almost any network.
In a perfect world with perfect networks, bandwidth is really the only thing that matters. A healthy, functional IP network that performs to spec will only show dysfunction when it runs out of capacity -- congestion initially leads to jitter and then eventually to loss. In this case, adding additional capacity is the appropriate answer. But this is rare in the world of network problems (see 'The bandwidth scam'). It's like the oft-mentioned belief that money makes you happy -- it seems obvious but apparently is never true.
The latest truism offers that quality of service (QoS) rationalizes the available bandwidth according to the needs of applications. We are given to understand that somehow QoS optimizes the use of bandwidth in support of optimal application performance. But careful consideration reveals that QoS mechanisms are typically some form of selective bandwidth reduction -- some applications see less bandwidth so that other applications can see … less traffic! There is no way that QoS can functionally increase overall bandwidth. In fact, QoS is more about putting bandwidth at a premium, both economically and in terms of availability, than about increasing access (see 'The QoS scam').
Understanding bandwidth is essential to a happy network and happy network users. Bandwidth itself is not an answer – but paying attention to it helps generate well-formed questions. And that's where good engineering begins.
About the author: Chief Scientist for Apparent Networks, Loki Jorgenson, PhD, has been active in computation, physics and mathematics, scientific visualization, and simulation for over 18 years. Trained in computational physics at Queen's and McGill universities, he has published in areas as diverse as philosophy, graphics, educational technologies, statistical mechanics, logic and number theory. Also, he acts as Adjunct Professor of Mathematics at Simon Fraser University where he co-founded the Center for Experimental and Constructive Mathematics (CECM). He has headed research in numerous academic projects from high-performance computing to digital publishing, working closely with private sector partners and government. At Apparent Networks Inc., Jorgenson leads network research in high performance, wireless, VoIP and other application performance, typically through practical collaboration with academic organizations and other thought leaders such as BCnet, Texas A&M, CANARIE, and Internet2. www.apparentnetworks.com