Is today’s computing paradigm, which has evolved over decades from the single central processing unit to networked or virtualised processing units, on which ‘applications’ are loaded and interact with ‘files’, but otherwise remained basically the same, adequate for tomorrow’s computing requirements?

Visions of ubiquitous computing and ambient computers have been with us for quite a while, mostly created by companies who would like to sell us the wall that turns into a large display, the coffee table that becomes an interaction device, the clock that announces important news, the mobile device with access to a slew of functions. 

Imagine, however, your grandmother trying to install that coffee table: “Please select the SSID for your network”, “Please assign a unique name to this table”, “Please select your server from the list of available servers”, “Please select your proxy settings”, “Select the applications that will run on this device” ... 

Clearly today’s network architecture is not ideal for handling all the devices around us that will be always ‘on’ and will interact with us and with other devices.  What is needed is an environment where there are no more servers and clients, no more applications that start and stop, no more files or documents that are saved, but rather a continuum where objects just exist and interact and are modified according to the environment, where data is not ‘saved’ but rather ‘exists.’ Each device is part of a whole doing things according to its own capabilities.  A ‘matrix’, if you will, just like in the film, where a world just exists and evolves.

For this to happen, several conditions need to be met:

- Data has to exist on the network, and not on one particular device or cluster. Google’s implementations have been impressive in this area.  Once, a whole Google data centre burnt down and users didn’t even noticed.

- Devices of all sorts (from the powerful laptop to the mobile phone to the clock on the wall) have to identify themselves to the network, communicate their inherent capabilities, and run the same functions at different levels according to what they are able to do.

- ‘Applications’, just like data, have to exist on the network and not ‘run’ on a specific processor. They have to split themselves between the powerful servers on the network and the user devices according to device capabilities and availability.  An ‘application’ has to be available from any device and has to be able to switch from one to another without any problems.

Now your grandmother can start writing you a letter on her desktop computer, and then move to the lounge where she can visualise and slightly modify the same letter on the display of the coffee table, without worrying about where she ‘saved’ the letter and without telling the coffee table how to connect to the rest of the network.

Science fiction?  Not really. A very exciting start-up (and Ariadne Capital’s newest portfolio company) by the name of GravityZoo has designed the architecture that will make this new paradigm a reality on the internet. With GravityZoo it is possible have an application ‘exist’ on a huge cluster of computers, and it is possible to access it from a PC at one moment, another PC in the next moment, and a mobile phone in the next.  Imagine, as a realistic scenario, that OpenOffice is “GravityZooed;” it becomes a suite of applications that are always available, with various levels of functionality depending on the capability of the access device.

Isn’t this just a web operating system?  Not really.  We are not talking about a complex application running at a data centre and a browser accessing it.  If everything was possible with a browser, surely Google Earth, Skype or Second Life would not force you to download complex applications in order to use their services.  This is technology that makes individual devices and servers irrelevant.  This links the whole network together into a ‘Matrix’.

By Julie Meyer and Jem Eskenazi, Ariadne Capital