Fifty years from now, some MBA candidate will look at the evolution of the Internet and have a pretty good idea whether the advances and evolutions of technology were part of some grand design foreseen by incredibly bright engineers and planners or if a bunch of reasonably clever people threw a whole bunch of technology against the wall to see what stuck.
A few years ago, companies came up with the idea of caching data in a distributed network so that the most commonly used content was stored nearest the user. Five thousand requests for an update of NFL scores made on a Sunday afternoon in San Francisco, for instance, didn't all have to travel all the way to the east coast if servers in the west were updated. All those requests were handled out there, cutting the per-update bandwidth expenditure to a minimum.
Regardless, Internet traffic has exploded, both because more people used the Internet, each used it more, and what they used it for migrated from static pages to bandwidth-hogs such as video.
Cisco, whose enterprise business, this GigaOm story says, has "pretty much stalled," said this week that it is introducing medianets. These entities will be optimized in several ways for the onslaught of video and other demanding content that the company hopes will continue to grow. The piece says medianets will cut across the vendor's consumer, service provider and enterprise business lines. The organized effort will extend caching capabilities to a new line of routers and position the company to establish CDN scenarios with customers, particularly service providers.
The success of this and similar plans-whether or not they will stick to the wall -- is predicated on the continued growth of high-octane content. A bit of good news-from Cisco's and other equipment vendors' viewpoints-came this week when QuickPlay Media released results of its first quarterly analysis. The results tallied statistics from 15 services run by the company. The results were that video-both video on demand and linear television-increased 87 percent over the previous quarter, downloads increased by 42 percent to an average of four per month, and average streams per user increased 12 percent to 19.3 per user per month.
There are other signs that demand is growing exponentially. The main point of this commentary is that there needs to be a new type of standard for online video. The important point is exactly why the writer is making that call: Previously, he says, most online video was produced by companies that made big commitments and sunk millions of dollars into projects. They produced a lot, but there were relatively few of them. The new generation of video producers are doing so as a far more mundane element of a holistic marketing strategy. There is more variation in size-there are small and medium-size companies in the mix-and, cumulatively, a lot more is being produced.
Cisco's move and anecdotal evidence from around the Internet suggests that capacity is rapidly filling. Of course, that's been the subject of a spirited debate among analysts. One side says that we are heading toward an Internet winter. The other -- while acknowledging that the pipes are filling -- suggests that the doomsayers are overstating the situation.
Nemertes Research president and senior founding partner Johna Till Johnson used her Network World column to reiterate her firm's belief that the access portion of the network will hit a wall sometime in 2012. Interestingly, the firm predicated its prediction on the idea that the natural creativity of humans will overfill the pipes. Just as people who invoke Moore's Law to predict the continued capacity growth of chips don't have to outline the technology behind every advance, it is possible to confidently say that demand will escalate without saying precisely which services will cause that growth.
Some people think that the Internet is heading to a crisis. Some are not so sure. Big equipment vendors are hedging a lot on there being at least enough congestion to make enterprises and service providers invest in hardware and software capable of alleviating backups. Who knows whose take is closest to the reality -- except perhaps an MBA candidate of the future.