Computers

IBM's MessageSight pulls data from up to a million devices

IBM's MessageSight pulls data from up to a million devices
IBM MessageSight: a mega-platform for the internet of things (Image: NASA Goddard Space Flight Center)
IBM MessageSight: a mega-platform for the internet of things (Image: NASA Goddard Space Flight Center)
View 1 Image
IBM MessageSight: a mega-platform for the internet of things (Image: NASA Goddard Space Flight Center)
1/1
IBM MessageSight: a mega-platform for the internet of things (Image: NASA Goddard Space Flight Center)

Three years ago, Google's Eric Schmidt announced that every two days, more information is created than was the case from the dawn of humanity up to 2003. According to IMS Research, by 2020 web-connected devices will create 2.5 quintillion bytes of information every day, with 22 billion internet of things devices up-belching information to the web. To marshal all that data, IBM has come up with a platform it calls MessageSight, which will allow any one organization to pool information from up to a million sensors and devices, at a rate of 13 million messages per second.

If this sounds detrimental to the signal-to-noise ratio of the web, consider that this data will be largely hidden from view, and potentially useful. How the data is put to use is another question, and with MessageSight, IBM's emphasis is on business. "This enables large volumes of events to be processed in near real time," says IBM, "allowing organizations to consolidate all of the information in one place and more easily glean insights to make better business decisions." Makes sense.

Perhaps more telling is IBM's illustrative example of how this might work. Here it turns to the automotive industry, which is a natural choice given the wealth of data modern cars are able to gather. IBM puts it like this:

"... an automotive manufacturer can use IBM MessageSight to help manage the features and services of its automobiles. With thousands of sensors in each car, a dealer can now be notified when a 'check engine' light turns on in a specific car. Based on the information transmitted by the engine sensor, the dealer could then notify the owner that there is a critical problem and they should get their car serviced immediately."

Though it's possible to read too much into what is an illustrative example, it's nevertheless suggestive of a corporate take on the internet of things. There's no reason at all why smart, connected devices shouldn't empower individuals and businesses, so long as they do not simultaneously disempower their customers. It's presumably the motorist who has paid for those thousands of sensors, and so it seems reasonable that the car should do its reporting to the owner, and not the local dealership.

Still, hopefully this is just an underwhelming example of the potential big data has to improve the lives of everyone, and not merely encourage us to spend more money.

Source: IBM

6 comments
6 comments
Snake Oil Baron
I can't help but think this would be very useful for science and ordinary people but specific applications are hard to dream up at the moment. Certainly it would be very synergistic with "smart dust" but the question of what applications it would enable still remains. The ability to gather and organize huge amounts of data is so new that our ability to know what data to gather and what to do with it when it is organized makes planning difficult.
Maybe having fine-scale, real-time magnetic field data for large areas would be helpful to geologists. Or help find plane crashes in mountainous wilderness quicker than search planes (I would guess that big chunks of metal suddenly appearing in the Amazon forest would register on a good sensor network).
Daishi
There is probably a use for auto manufacturers to monitor how people use their cars (mileage, average speeds, actual MPG, how well some of their engineering seems to be holding up etc.). Electric car manufacturers could monitor how well their batteries are performing in different climates (hot, cold) under different styles of driving. Just have the cars push the data via 3G or WiFi after you pull into the driveway.
Statistics like this is part of what allows F1 teams to succeed and I'm sure it could help auto manufacturers refine their designs or at least provide them a bunch of real world numbers to go on.
I am sure as the internet of things gets more popular there will be greater need to sort the data. As the companies that know the most about you are getting rewarded with the best advertising dollars what could possibly go wrong?
ie "We couldn't be happier with our Google Household Robot (tm) but once we got it we noticed all the commercials on TV were for romantic getaways. After a while all the commercials started trying to sell me male enhancement drugs and now they are trying to sell her a personal trainer and sex toys. Should we consider moving its charging station out of the bedroom?"
Michael Woolf
Wow that is fake as hell. Why do all these light maps put so many dots in australia where no one even lives.
(The image includes fires, which is explained in the image source - Ed.)
f8lee
Is this what the NSA uses to run Echelon?
Murgatroydals
Murgatroyd will employ a system of cognition utilizing a concept such as this.
VK3CRG
I tend to agree with Michael Woolf to a degree.
Something is very wrong with the data for Australia.
All those dots on the NW part of Australia are in the desert.
There are no people (practically none!) or "fires" out there. There is nothing to burn! It's just thousands of kilometres of scorched red earth, yet it covered in dots on the map.
Oddly enough major centres like Melbourne & Sydney (about 10 million people between them!) both just have one dot each.
Something is very wrong there somewhere in the Australian part of the map anyway.