Two of the most exciting technology companies these days are Amazon.com and Sun. They’re interesting because they’ve identified an impending sea-change in the computing landscape, and positioned themselves to take advantage of it. That makes Wall Street and the public skeptical. They much prefer a narrow focus on the next quarter and extending existing markets. Yet such linear thinking is what produces the innovator’s dilemma trap identified by Clay Christiansen. If the world is changing, you can’t thrive by standing still.
And the technology world is changing.
It’s changing through the emergence of what I like to call Internet-scale computing, which sometimes goes by names such as grid computing and utility computing. In this model, applications run across a distributed, fluid, globally accessible fabric of resources, rather than a local machine or a remote server farm.
Internet-scale computing is broader than the Web. The basic idea is that, to quote Sun’s famous tagline, the network is the computer. Sun has used the expression for years as a metaphor, but now there’s enough connectivity, with enough standardization, and cheap enough storage and processing to make it literally happen.
The company most deeply committed to the Internet-scale computing paradigm is Google. That’s why it’s building football-field sized data centers stocked with vast numbers of cheap computers running its homebrew software. Microsoft, with the arrival of Ray Ozzie, has also gotten networked computing religion, although it has a long way to go to shed its anchor of desktop software. IBM and HP seem more focused on commoditizing computing to move the value up into services, which isn’t quite the same thing, although both have significant investments in Internet-scale computing. And don’t ignore eBay, which may become the NYSE of the Internet economy, the vast transaction-processing engine of an inconceivably massive market.
All those comp[anies, however, see Internet-scale computing primarily as a way to deliver their own services. Which brings us back to Amazon and Sun, which are taking more of a platform approach. Amazon just announced of Amazon’s EC2 utility computing initiative, on the heals of its S3 virtual storage offering. Sun has a similar offering at network.com. Amazon and Sun are different companies in almost every respect, yet they’ve both identified a major opportunity in building scalable infrastructure for Internet-scale computing.
I admit to some bias, as both Sun CEO Jonathan Schwartz and Amazon.com CTO Werner Vogels spoke at Supernova this year. And having the right vision is no guarantee of success without great execution. That being said, both these guys very deeply get it. I don’t think the market understands what they are doing, but the wind of technological change is at their backs.
So, why did I call this post “the new edge and the new core?” Think about it. We all like to think that the world is decentralizing. That’s a key theme of my writings, and of Supernova. I’m still convinced it’s true, but we need a better understanding of what decentralization means.
The Internet-scale computer is the network; it’s the core. Sure, it’s an open, flexible core connected to smart edge devices, but it’s the core nonetheless. Call it a decentralized core.
What, then, is the new edge? It’s what surrounds those football-field Googleputers in Oregon, and all the other manifestations of the Internet-scale computer. It’s our friends, the network operators. The telcos and cablecos own the pathways radiating out from the brains of the network to what lies beyond… which is all of us.
The real battles among all these players have just begun. It took twenty years from the adoption of TCP/IP to build out the commercial Internet. It took ten more to deploy the always-on broadband network we have today. At the same accelerating rate of change, give it five more to finish building the Internet-scale computer. If you’re playing this game on any shorter time scale, you’re in trouble.