To: FJB who wrote (34723 ) 7/31/2010 7:32:29 PM From: Frank A. Coluccio 1 Recommendation Read Replies (1) | Respond to of 46821 Hello FUBHO, WLD, All: Before classifying latency and bandwidth as two different species, I would note instead that latency and bandwidth are perhaps two of the most widely misunderstood attributes in the field of transmission, and that goes for practitioners as well as lay folk. What is bandwidth? What is latency, and what are its implications? Most popular interpretations of bandwidth today equate it with a channel's bit rate, irrespective of the "width" of the "band" of frequencies being utilized by the underlying channel (the medium). In fact, a lower bandwidth channel can indeed deliver a higher bit rate if the former uses a suitable encoding scheme that results in a higher bits per hertz efficiency. Likewise, the popular notion about latency fails to take into account its many manifestations in the time required to deliver a single bit, as opposed to a frame, or a packet, or a complete block of information before any meaningful use could be obtained from it. I would like to clarify one point I made, half in jest earlier on, regarding the use of clouds as platforms for online gaming applications. As noted, it was only half in jest. I was not referring to the suitability of a public cloud to necessarily perform central-site processing of gaming applications and rendering, per se, although that remains a possibility too. Instead, what I was thinking about in a half off-hand way was the potentiality of the larger cloud entities -- not social nets, but data center outsourcing entities on the level that provides storage and cloud services for enterprises, say, such as Amazon's EC2, Microsoft's Azure, and the like. These entities could conceivably move into adjacent market areas and begin offering piggy-back transport services along their established dedicated routes, as a means of performing an "Internet bypass" function, for applications such as gaming that require low latency, since the cloud entities I've mentioned have already taken measures to assure the highest possible latency performance across those routes for their own purposes. Yes, it's probably a wild-arsed idea, granted, but that's what I was referring to earlier, fwiw. FAC ------