The initial challenge for the Internet of Things (IoT) was how to provide physical connectivity of small and often remote devices to the Internet. This issue has basically been solved with the plethora of wireless connectivity solutions. The real challenge for IoT is data organization, sharing and search on an unprecedented scale.
Most discussions about IoT usually point out in passing that today’s machine data is concentrated in vertical, or isolated, data silos. However, they rarely explain why this is a problem. More important, they do not give a roadmap of how to solve this problem going forward. I will try to provide some insight on both points.
So, what is the big problem with data silos?
Let’s parse the data silo issue with a practical example based on the transport industry. It may come as a surprise to you, but you have actually been living in quite a few pre-IoT (I prefer to differentiate them as M2M) silos for most of your life. This is most evident in how our transport industries work around the world.
+ Also on Network World: 6 tips for working with IoT and big data +
The transport industry is of course the industry that builds and manages the roads and rails that we all ride as part of our daily business. What is not so obvious is the massive amount of information and communications technology (ICT) that underpins and supports the functioning of all this infrastructure. This comes in the form of vast deployments of “things” such as Bluetooth traffic flow sensors, parking sensors, Automatic Number Plate Recognition (ANPR) cameras and CCTV systems, to name only a few.
These “things” produce large amounts of data. The problem with this data is that the only place it goes is back to a command and control system to be used for important, but very limited, purposes. Now compound this with the reality that transport systems in any country are typically independently managed under regional transport authorities. For example, in the U.K. there are over 100 of these! The result is not so much a silo problem but a silo of silos problem.
In this context the problem is clear. The data generated is vastly underutilized and is closed to the outside world, stifling further innovation. More than that, because each regional authority does things just a little bit differently, they are unable to easily share data between themselves never mind another vertical industry.
In my opinion, it is in this next level of data sharing and data fusion that we are likely to see some of our greatest leaps forward. Stated more tangibly: Perhaps it is quite common today for a regional transport authority to commission development of user applications. The problem with these apps is they work only in one specific region. This is all the more ridiculous when you realize regions may only be separated by the width of road or a backyard in most instances.
oneM2M and other web services exist to break down silos
The good news is that we have a clear path for solving the IoT data silo issue by extending the current web services approach from the internet.
Web services, of course, refers to using standardized protocols such as HTTP to communicate between computers on the internet to perform a useful service. For example, in a business-to-business (B2B) banking application, a computer from one bank may electronically transfer funds to another bank at the end of each business day to settle all outstanding transactions.
As mentioned previously, web services require a standardized communication protocol, such as HTTP. They also require a standardized data format, such as XML, to exchange data. The final critical component is easy-to-use application programmer interfaces (APIs) so that programmers can easily invoke the web services from inside their software.
Now let’s apply the web services model to IoT and see how it can help break down the IoT data silo isolation problem. Fortunately, we already have an advanced solution available through the oneM2M international standards body that we can use as an example.
oneM2M is a web service-based architecture specifically designed to create a horizontal service layer that can span different business domains such as transportation, utilities, and healthcare. Thus it is targeted to breaking down the IoT vertical data silo problem. It does this by using protocols such as HTTP and the newer Constrained Application Protocol (CoAP) protocol for communication. It also defines detailed data structures, semantics and APIs so that third-party application programs can easily use oneM2M standards to exchange IoT data between different service providers.
If we applied the oneM2M solution to our previous example, the solution would go something like this: By incorporating the oneM2M APIs into future transportation products, the impact would be nothing less than profound. In one step, it would put the transport industry on the next-generation roadmap to the true Internet of Things. This would allow the transport industry to control and share their data in a disciplined manner between themselves, third-party application developers and other industries. Applications developed in one region would now work everywhere.
Do we still need search engines?
You may at this point be wondering if we still need web search engines as per my introductory supposition if we have horizontal solutions like oneM2M to organize and exchange data between applications. The answer is a very big yes. We will still need search engines. This is simply because the data problem in the IoT is almost inconceivably massive. This challenge dwarves the data management challenge that was the World Wide Web.
The roadmap to the IoT will be defined by progressive data organizations and alignments, not all happening at the same pace or priorities. In this, I believe the oneM2M standard will be very important. However, it will never cover all IoT deployments. Just as for example, today there is no single B2B solution that satisfies all businesses. I expect that alongside oneM2M there will be other standardized and proprietary web service-based IoT solutions, and this will be quite healthy.
The end game of the IoT will not be an elimination of silos. Rather, it will be a consolidation of them to a manageable level. It is at this point where we will need our next-generation search engines to make sense of it.