Stay on Top of Enterprise Technology Trends
Get updates impacting your industry from our GigaOm Research Community
[qi:051] Mark Cuban, the histrionics-prone billionaire and owner of the Dallas Mavericks, caused a major ruckus over the weekend when he wrote a stinging essay entitled “The Internet is Dead and Boring.”
The essay lit a fire under the bloggers, who turned into NBA officials for the day, declaring – loudly — that Cuban was in the wrong. But his arguments — while laced with impatience and full of disregard for the financial realities faced by incumbent carriers — when viewed through an infrastructure lens are in fact quite sound.
…. if you index the expected growth in bandwidth consumption by applications that are heavy LAST MILE bandwidth users (as opposed to the Internet backbone where there is plenty of bandwidth but consumers cant get to it) vs the actual increase in LAST MILE bandwidth available to the home, our net effective throughput to the home could decline over the next few years.
Here Cuban is spot on: Even as the connections to our home are becoming faster, the carrier networks (including the last-mile connections) are getting filled at an even faster rate. The reasons for that are manifold, with video being the most obvious candidate.
But the bigger question Cuban is asking is whether the Internet’s infrastructure is sufficient to keep the innovation cycle moving forward. And the answer is no.
I have often argued that just as gigahertz drove the WinTel ecosystem, bandwidth (both up-and downstream) drives innovation in the broadband world. Of course, it would be too simplistic to correlate last-mile bandwidth with innovation, for the network architecture is inherently complex, and is getting more complex by the day.
It was during a recent conversation with Rohit Sharma that it all came together. Sharma in his past life started ONI Systems (an optical systems company that was acquired by Ciena Systems (CIEN)) and is now a partner at Mohr Davidow Ventures. Over cups of steaming espressos, we reminisced about the optical bubble.
We couldn’t help but compare our tallies of ex-telecom executives that are involved in either media- or video-related companies instead of next-generation infrastructure. It quickly became clear to both of us that front end was getting all of the attention while the back end was being all but ignored.
Is today’s infrastructure adequate? Given that there will be more than 2 billion phones and a billion PCs connecting to the “network” in 2008, the back-end infrastructure is going to need a rethink. We’ve been going on the same set of core technologies that made their debut in the late 1990s. Cuban isn’t the only one arguing for a rethinking of the Internet as we know it. From the members of the networking consortium known as Internet2 to the Japanese Communications Minister Yoshihide Saga, there is a burgeoning conversation taking place about developing a new and improved version of the Internet.
I asked Akamai (AKAM) CTO Mike Afregan about Cuban’s assertion and the infrastructure challenges that loom ahead. “Look, we have had the same transport protocols and routing protocols, and they are not going to change,” he said. “But I think it is clear that the nature of the Internet is changing, and there is so much work that needs to be done — from rethinking chips to better storage systems to disks to systems.”
John Roese, chief technology officer of Nortel (NT), is of the same school of thought. “If you look at the progress made from 300-baud modems to 10-Gigabit Ethernet,” said Roese, “the cost per bit has declined by a factor of 22 million to one. But that isn’t reflected in the consumer Internet experience.”
The fault is in the asymmetric nature of the Internet. The downstream speeds are getting higher, but upstream speeds are still being controlled in a miserly fashion by ISPs, thus acting as a break for truly interactive applications.
An asymmetric Internet is good for disseminating information – after all pulling down information (or YouTube videos) moves packets in one direction. This is perhaps the point Cuban is trying to make when he says that the Internet is like a utility and therefore boring. Electricity, after-all, also works as a one-way service — it comes into our house and we use it for everything from stereos and air conditioners. Today’s Web and Internet applications are doing precisely the same on our desktops.
The future, however, is in two-way, symmetrical Internet, where applications such as Kyte.TV and Sling Media can actually be put to use. “The Internet needs to become interactive, bidirectional, and be contextually aware,” argued Roese.
Roese agreed that Internet in its current manifestation is a lot less exciting, though he was more measured than the bombastic Cuban. “What makes me nervous is the lack of startup activity around the next-generation Internet,” he said.