| Tech and Business Videos |
|
|
|
| Gridlock looms in cyberspace |
March 17, 2008 |
 But a new marshal may be riding into town
By Ian Harvey
Is the Internet about to stumble under the burden of its own success?
Illinois-based Nemertes Research Group raised eyebrows last fall when a report it wrote declared the Web could slow to a crawl in two years, clogged by traffic. Nemertes said up to US$55 billion in infrastructure investment is needed by 2010 — about 70 per cent more than has been budgeted by the infrastructure companies. And that position is echoed by market research group TeleGeography, which noted traffic was up 75 per cent last year while capacity grew only 47 per cent.
“Failing to make that investment will not cause the Internet to collapse,” the report stresses. “But it will throttle innovation. The next Google, YouTube or Amazon might not arise.”
Apparently it’s our fault: “Online video demand in North America jumped from seven per cent of traffic in 2005 to 18 per cent of traffic this year, and it will grow by a factor of 19 through 2011,” said Olaf Krahmer, Cisco Canada vice-president of service provider operations. Add Internet Protocol TV streaming, VoIP, peer sharing and more Web-based applications, and the road is looking even more congested.
Web traffic cops
Part of the cure may prove bitter medicine for those who embrace the Web’s Wild West mentality where anything goes to anyone at anytime in any place. That’s because Internet Service Providers figure they can coax capacity without investing heavily in new fibre by managing — some say massaging — existing traffic to speed things up and slow other things down. This is where it gets sticky. By playing traffic cop and threatening tolls for “premium data,” netizens say ISPs are striking at net neutrality — the concept that the Web is free for all to use without fear, favour or censorship.
That ship, however, may have already sailed. Rogers Cable, for example, “shapes” traffic on its network, inspecting and identifying data packets and slowing some down if it decides they are lower priority. But that’s not as nefarious as it might seem, said Tom Donnelly of Sandvine, a vendor of such traffic shaping technology. Rules have to be made for the greater good, something we didn’t need in the carefree days of Web 1.0. If 20 per cent of subscribers are generating 80 per cent of the data traffic, why should the majority suffer a loss in the quality of their service?
“It’s the tragedy of the commons,” Donnelly said. “As an ISP I’d say you’re obligated to make intelligent decisions to ensure the efficiency of the network.” Shaping, Donnelly said, isn’t censorship but configuring traffic based on user needs, bandwidth demand and the type of application the data represents.
VoIP is sensitive to lag so it must get priority; e-mail can take a couple of seconds longer; that YouTube funny may spool up slower and that feature film you downloaded may take five minutes longer, but if all traffic were treated equally during peak times, there could be congestive chaos and a drop in quality of service. “Some packets are motorcycles so you want to get them in a lane, others are transport trucks and you want them in their own lane,” Donnelly said.
|
The next Google,
YouTube or Amazon
might not arise. |
|
| Or new tech will save us
There is some hope, though, that technology itself may avert the overly stringent shaping and prioritization of packets. ’Net giants like Cisco and Nortel are pushing solutions that wring more speed and capacity out of the existing infrastructure in the same way hot rodders squeezed horsepower from the family sedan. This may mean 8Mbps broadband jumps to the equivalent of 100Mbps while other encoding drops the bandwidth requirements for HD TV to 8Mbps from the high teens.
But just when you think you’re out of the woods, along comes the mobile Internet, said Cisco’s Krahmer, which itself is expected to grow by “a factor of 17 through 2011,” adding a whole new traffic burden.
It all adds up to a huge headache— which in turn means there’s an opportunity for someone to make money. The question, though, is who will pay: the consumer or those enterprises that depend on high bandwidth to reach their customers? History indicates it will be the consumer who pays.
ExecutiveOverview Archive
|
|
|
|
