The claim for being first with networked embedded control devices helongs to Echelon [http://www.echelon.com/company/] - dating some 20 years. Efforts to assimilate embedded devices into the Internet are futile! :-) [as the market fragmentation demonstrations.]
Hi Rick, we wrote the white paper behind the article *precisely* because most folks focus on the "things" - the endpoints - endpoint development environments (Arduino, Berg), endpoint wireless connectivity solutions ( Broadcom and NXP...the sensor equivalent of broadband's famed "last mile"), endpoint security, endpoint sensors, etc. [hopefully Roger likes my use of "endpoint" all over this sentence].
When you envision IoT futures, those "things" must be connected somehow to the other hot topic, Big Data, the cloud-based analytics that generate the contextual smarts for new ambient and wearable systems. But the connective infrastructure is often left out of the discussion.
We didn't miss these other activites, we deliberately side-stepped them to focus a portion of the discussion on larger organizations and companies who are taking the first steps toward building the pervasive connective infrastructure that affordable services will depend on to enable smaller, more power efficient wearable and embedded tech to be smarter than their power and price budgets would otherwise allow.
Please click through Pat's article to our paper. And then let's talk about mapping out the large scale features of an ambient computing future. End-to-end solutions must have a middle... ;v)
It is true that the term IoT englobes many technologies and interpretations. In a way it reflects the exciting realm of possibilities we now have. I worry about security issues to be honest and I think many are under-estimating this real problem.
While reading your post I flashed back to the late '80s and early '90s when I was explaining to confused executives what this new "Internet" thingie was and what it meant to their businesses. This iteration is even more far-reaching, but the basic principles of distributed sensing and computation apply. IoT is in one respect an advancing wave of connected sensors and smart devices, but it is also an incremental expansion of the Internet. Rather than replacing what has gone before it will force it to evolve. Big data analysis will need to be able to handle truly massive amounts of data in real time. Networks will need to handle cross-flows peer to peer and outside-in data flows as effectively as they currently handle data flows that are primarily from servers to the edge of the network.
Business model changes are going to be interesting to watch as well. I heard a network equipment exec at one conference talking about how they were enabling an ISP business model that he referred to, without any trace of irony, as being a "Two-faced business model". The idea was that the ISP could trace the origin of data flows to enable them to charge Netflix, Google, etc. based on how much data they originated in addition to charging their customers to access those flows. How does this change when Google is aggregating data flows from millions of sensors?
It seems to me that part of the problem with IoT at this early stage is that the concept is not totally defined. I still think IoT means different things to different people. It's early days for IoT, but there is a lot of work needed in standardizing on technologies.
>> The idea was that the ISP could trace the origin of data flows to enable them to charge Netflix, Google, etc. based on how much data they originated in addition to charging their customers to access those flows
That will be the battle of the next decade. The way my internet connection slows down when the college kid that stays closer by returns to when he is gone demonstrates that Netflix and co may have to support the pipes that keep them in business. If they do not and the network owners refuse to innovate, their business model will be in jeopardy.
>> Efforts to assimilate embedded devices into the Internet are futile!
Why do you think so? I generally think that the business of Internet assimilation is straightforward with the IP being the key part. If the device has the IP defined, it can get it. That is different from making money in a crowded business but technically it is not hard. What is that a futile endeavor?
There are many aspects to this IoT, so naturally you're going to have lots of organizations scrambling in lots of directions. I think the biggest failing in this thinking, though, is the notion the the IoT is something new. It's not. It's nothing more than continued evolution of the Internet. Gradually more and more of the same.
There's a very good book on Internet Protocols called "Internetworking with TCP/IP," by Douglas E. Comer. My edition is dated 1991. Up front, it states that the Internet essentially started in 1980, and had grown to "hundreds" of individual networks, and 20,000 computers, by 1987. By 1990, that had grown to 3000 networks and 200,000 computers.
Surely, where we stand today would have been considered "the Internet of things," with a 1991 perspective. Used to be that access to the Internet was only via remote terminals connecting to large mainframes, via some sort of telephone or other connection. Wouldn't direct IP connections to PCs, tablets, smartphones, sensors, machinery, all of which exist today, have been aptly called "Internet of Things"?