We’ve heard so much about the Internet of Things (IOT) being the next enormous step for both the Internet and for Big Data. For those who don’t know, the IOT is about getting devices and sensors (things) connected to the rest of the world in much the same way people are today through the Web and smartphones. Gartner claims the IOT trend will add 50 billion connected devices by 2015 on top of the 2 billion connected people who currently populate the Internet as we know it.
In a story on TechCrunch today, the landscape of the IOT was laid out in all of its glory. This is the new playground of the Silicon Valley (and elsewhere) VC and entrepreneur and is attracting lots of attention and cash. But is the tip of the spear where the real money will be spent to empower the IOT?
Wireless on steroids
What really powers (in theory) this hodgepodge of software and hardware is wireless technology that links the far-flung sensor with the enigmatic cloud (which maybe far-flung, maybe next door…it’s the cloud, who knows?). If 2 billion people struggle to be connected today, the need to add 25X devices is going to require a whole new way of connecting things. It will take every innovative idea and technological breakthrough we can muster.
It will require wireless capabilities that don’t currently exist.
Streaming technology
Even more, the IOT is going to require an ability to catch and make sense of streams of data that are generated by all of those billions of devices and sensors. This isn’t a database problem to solve, either, as there’s no time to store and then retrieve data coming form the other 48 billion ‘things’ much less the stuff created by human beings. This will require technology built to absorb these streams and make sense of them as they arrive, at full velocity, at the point of collection.
Those streams will need to be correlated for patterns that show opportunity and risk. When the moment is just right, rules will need to be applied to kick off messages to people or other systems. This is a process bonanza that will require real technology muscle to be flexible and fast enough.
While this is done in the financial services markets today, this technology will need to be applied to supply chains, logistics, healthcare and many more business models.
In-memory storage
If it can’t be stored in a database, information streaming in from the IOT has to ‘live’ somewhere, and the only place fast enough and nimble enough for high-speed information is in-memory storage. Think: flash memory at extreme scale. Without in-memory computing power, the IOT would be powered by traditional technology architectures and that will never work for the speed and scale of the problem.
Integration
While wireless technology will carry it, streaming technology will ‘read’ it as it arrives, and in-memory storage will hold it for just the right amount of time, it all lacks value without integration with existing information and systems that can put things in context and help people make business and personal decisions. As much as technology captures our imagination, it captures the pocketbook only when it fits a business purpose and generates revenue. That will happen not when TechCrunch creates the VC and entrepreneur’s mashup above, but when business and individuals derive value from it.
On this front, much of the technology to integrate exists today but isn’t broadly adopted yet. What was once the luxury of high-speed business models will need to be the heavy lifting infrastructure of companies that want to compete on a level playing field. This is exactly why Gartner believes that integration will be 50% of the challenge of major projects for the next several years.
The Internet of Things is a fascinating topic that will absolutely change our world. Let’s get real about the technologies that will power it.
Nice write up Jeanne. One point worth noting on the idea that “it can’t be stored in a database”. That isn’t quite right - much of this will be stored in a database (specialized like Informix Timeseries or a platform like Hadoop or Cassandra) but it will be handled/reduced and put to use before it gets there. That is a pretty well established pattern already with the big telcos.
The other key point here is that the best Streaming technologies do the integration on the fly, often from many disparate sources before it ever lands anywhere. That challenges the notions that many have over what/where/when integration happens.