Cosm (formerly Pachube) founder, Usman Haque, tells Ian Scales how the London-based user-driven Internet of Things platform is about sharing data and creating 'cosms'.
Just for the record, 'Pachube' was a derivation of patch-bay. This was the early screw-driver and soldering iron idea of the internet of things as a whole lot of connectivity going on. Now, says founder and CEO of Pachube, now Cosm, Usman Haque, that idea had become less useful. The site/community is all about shared 'workspaces' and 'environments' ('microcosms' and 'macrocosms'), so Cosm was born.
Cosm was started by Usman as the Pachube prototype back in 2004/2005. He had trained as an architect and, he said, had ended up doing a lot of work on interactive environments. "I developed it as a generalised environment to handle data from sensors: thinking about a patch-bay to enable anything to be plugged into anything else."
He spun it off as a separate company called Connective Environments in 2008 and it was acquired by LogMein last year. It's essentially a platform for connecting devices across the Internet and then making sense of the data collected, ideally in collaboration with people doing similar or complementary things on the same platform. Think MySpace for IoT but without the grizzled Australian proprietor.
Cosm is the natural progression of Pachube," says Usman Haque. "It's a broader idea and much more to do with handling more data and creating what we think of as 'cosms': microcosms or macrocosms of data and people and team members."
So how is it different from other IoT efforts, would he say? "First, there's a very strong focus on putting control in the hand of users. We strongly support open data, formats, make things as simple as possible for people to build and innovate. Secondly, we are borrowing techniques to scale massively. We're able to build on the technologies that have been introduced for things like Twitter, which has to handle billions of tweets. What distinguishes us is not just about handling data in the abstract: we are thinking about what people are going to do with that data; how they're going to make use of the data to make sense of their environment.
"The big difference with the Cosm approach and the 'old' approach is that where before technology was deployed to create closed systems where you'd deploy your sensors and devices and deploy a service for managing them and you would probably also control the service that was then sold on other customers and therefore to another business. What's different with IoT is that when you're deploying the products you don't actually know who's goind to extract value from them."
There is a palpable whiff of idealism hovering around the edges of Cosm. As you might expect with a 'bottom up' project which 'enables' people and organisations to build an Internet/platform of Things, it appeals to those who have what we might describe as a political or social 'passion' and are prepared to put time in to develop something which might assist the cause. It's therefore partly a space for techies who need a purpose beyond, say, building a bomb-proof sensing system for an industrial process.
Environmental concerns are very front and centre and perhaps its most successful and well-known application so far is AirQualityEgg, a global pollution sensing application.
Usman seems to be aware of the trap here. Building a community around technical openness and innovation applied to things like pollution control or sustainable smart cities attracts a committed bunch of community contributors and champions for your platform, but there's a danger that the platform can be seen as so worthy and 'not for profit' that it may seem less fit for commercial purposes. Usman is keen to stress that the platform open to all and very much open for business.
From the very start the data model itself has not just looked at a single device or object in isolation but has always looked at environments. An environment could be a home or it could be a particular room or it could be even just a mobile device that has multiple sensors in it. The point is that in all cases we're looking at things which quite probably have multiple parameters.
"We're mostly interested, frankly, in handling contextual data not just the parameter data streams, so it's important to know where things are, what the owner or author of that environment has said is important, and that is the wrapper for the the time series parameters which might actually be changing," says Usman.
So where is the IoT at present?
"We are still so much in the early days and I sometimes compare it to the days of the web when we browsed using grey backgrounds and blue underlining for links. But there's lots of clues and there's people starting to make money off them. Tthere are different ways of scaling up in the business sense, but what we're really looking for is for different companies and organisations who have specific domain knowledge about their particular industry to start to build specific IoT services."
Another important point is that Cosm is not about creating a data free-for-all but, he says, "the chances are that in making data available to others (carefully controlled) they'll help extract the value because they'll bring some kind of domain-specific knowledge to it. We sometimes describe Cosm as being like a data escrow service."
"One of the technical challenges of the legacy systems is that it's very hard to open up parts of it without incurring lots of security risks, whereas if you build the system from the beginning with the sharing intention then you can think about the security implications afterwards."
I pointed out that a lot of the participants seem to see the IoT as a way of democratising information.
"Because it starts from the ground up there are a lot of people who just start out wanting to monitor their own home or energy usage or even air quality. But, because they're able to make the measurements over time, they naturally start thinking, 'what happens if we can compare our energy use with each other?' It's about people making sense of their data in conjunction with other people."
And as for smart cities, these too have to be thought of around the other way.
"They're often described in terms of being big ultra-reliable transport systems and so on, designed from the top down. But these are actually phenomenally expensive. On the other hand if you think that what cities are all about is dealing with the unplanned and unexpected, then a smart city infrastructure might not be so much about efficiency but about enabling people to connect with each other or to innovate themselves.
"What we really want to do is make it really quick and easy to get going making stuff with devices - if you're an individual or an engineer working for a corporation, you so often have to send off for an SDK or you must have specific hardware. We just want it to be possible to have people sign up and then get going almost immediately. You sign up, you're working first with your own data, and then you're following other people's data, you start working with a team and working collaboratively on it."
Not so much like an IoT MySpace, then, more a Facebook.
please sign in to rate this article