The emergence of distributed peer computing popularly known as P2P (peer-to-peer) marks an important new phase in the evolution of the ASP model.
In the early days of the ASP industry, purists used to argue that the model was all about migrating computing away from local devices into massive datacenters deep in the heart of the Internet. But that argument has always been a fallacy. In today's ASP environment, there's plenty of evidence to show that computing will remain pervasive throughout the network, all the way out to the user devices on its periphery. Here are some of the most persuasive points and some important news about what comes next.
Imperfect connectivity
The pervasive network is a mirage. There will always be times when users are disconnected or poorly connected, even if only temporarily. At such times they'll need local processing to maintain acceptable computing performance. Paradoxically therefore, there is only way to truly fulfil the dream of anytime, anyplace access to computing the online environment must include support for intermittent offline operation.
Remote management
Computing can still be managed by ASPs, even if it's not in their datacenters. One of the hidden stories of the ASP industry is the remote management of computing left in situ on the customer site. Taking an example at random: This is the primary business model of many MSPs (management service providers) such as Silverback Technologies or McAfee.com companies that take responsibility for monitoring a customer's vital server clusters, or for keeping desktops free of viruses. They don't ever attempt to do this without installing some software agents or hardware devices of their own at the remote location. It's the only way they can guarantee a reliable, consistent and fully self-managed service.
Resource management
The other two were for starters; this is the killer argument. An established tenet of computing science is that processing migrates to where the data is. It's simply not an efficient use of resources to shuffle data around the network just so you can keep the smart computers all in one place especially not when you keep users waiting while you do it (they're the smartest computers of all).
So does that mean we shouldn't have been transferring all our applications and files out into ASP datacenters to start with? Thankfully, it doesn't. What we've been doing is taking applications and data that used to be kept inside the enterprise and migrating it out onto the Internet. That's happening because businesses and individuals now need to collaborate with each other outside of the enterprise, and the data for those collaborative operations needs to be available on the Internet.
But when an individual is creating new data or editing existing data in a new way, it doesn't make sense to process that data on a computer far away. Far better to process it locally and then deliver the results from the individual device out into the collaborative Internet.
Internet peers
Distributed peer computing supports this paradigm, but in a new, Internet-friendly way. Instead of dividing computers into those that are on the Internet and those that connect to it, it allows the Internet to reach out and embrace all computing, wherever it is.
One of the best examples is by Groove Networks, the startup founded by Ray Ozzie, who pioneered groupware in the late 80s as the creator of Lotus Notes. Called simply Groove, it allows users to expose data directly to other users while they're working on at their computers. When they're ready to go offline, they can store and publish the data on an Internet server, then continue working on it offline and resynchronise when they reconnect.
This is the start of the next big wave in Internet computing, probably the most important next step since the advent of the web-based application server. It's new and it will take maybe five or more years to start to become mature, but ASPs need to be watching it now, adjusting to this new mindset, and starting to build it into their offerings.
In doing so, some will likely evolve to become subscription computing providers, such as CenterBeam, the startup founded by Sheldon Laube, who at Price Waterhouse in the late 80s championed the first corporate adoption of Notes (coincidence?). These are companies that take responsibility for providing and managing the local computing resources and Internet access for business users at the network periphery. They will play a vital role in maintaining Internet computing infrastructures that will allow peer computing to flourish.
The advent of distributed peer technologies provides yet another illustration of how radically computing is reshaping itself as it joins to the Internet. ASPs must not only become adept at managing computing within their own premises on behalf of their clients; they are also going to have to learn to manage computing at arm's length, wherever it happens to be.