The recent launch of the Open Connectivity Foundation (OCF) was met first with a wave of “oh good, another IoT consortium”, then “phew, it’s just a rebrand of the OIC”, followed by a bit of confusion over why a few AllSeen Alliance players and some other names jumped in. Is it just a marketing ploy, or is there more to this?
Understanding the technology consortia game is a study in bringing order out of chaos. A big point in that is understanding the difference between a standard and a specification. To have a standard, ratification from an internationally-recognized standards body must be obtained – think IEEE, ISO, ITU, SAE, and others. These standards bodies are famous for processes in review, IP disclosure, balloting, commenting, and revision control. A breadth of technology is represented, usually in multiple categories. Efforts are not necessarily aligned with commercialization; many standards are developed and languish, never widely adopted.
Marketing consortia, on the other hand, are about specifications and driving commercialization. Their efforts may ultimately be submitted to a standards body for ratification, but unless the consortia leadership is experienced and disciplined converting a specification to a standard can be a daunting exercise. Organizations are usually a band of people representing companies with common interests, in contrast to standards bodies which are composed largely of individual subject matter experts, although most tech firms understand seeding those organizations with their talent is in their best interest. Marketing consortia typically develop some type of branding effort, and there are membership fees and usage rights and all those details.
Now that we don’t say standard when we mean specification, let’s look at what really happens. A well-run marketing consortia can move quickly and create a de facto usage model (see what I did there?). For example, Intel was the driving force behind the PCI-SIG and the USB Implementers Forum, but they brought in many other credible voices and did the right thing for the industry, not just themselves. If you can get enough people in an ecosystem to agree and demonstrate compliance, technical harmonization removes friction and buyers proceed with confidence in adoption. Competitors find that early-stage cooperation helps everyone, compared to trying to go it alone in the marketplace.
Those were subsystems, however. The problem gets a lot tougher at the system level, where a customer is faced with integrating disparate efforts. An excellent example of just how messy this gets was in the telecom industry a few years back. There were hardware organizations, such as PICMG and VITA. There was The Linux Foundation driving open source adoption. There were several organizations, including the Service Availability Forum and SCOPE, driving middleware specifications. Each was doing what it thought was the right thing for carrier-grade implementations – but when their specifications arrived at the carrier customers they were targeting, there were gaps or conflicts. In an attempt to facilitate cooperation and harmonization of competing specifications, the Mountain View Alliance was offered as a consortium-of-consortia.
IoT consortia have been looking like the same mess in a much bigger systems problem. About a year and a half ago I wrote a SemiWiki piece titled “Chip side of the Open Interconnect Consortium” describing the problem and introducing the OIC as a way to get the chip guys and the software guys talking. I pointed out that the AllSeen Alliance was missing support from chip vendors (other than Qualcomm themselves), and the OIC was backed by Intel and Samsung with Atmel and Broadcom tossed in. I said that Broadcom has a history of being “open source hostile” and sure enough, they made a rapid and noisy exit from the OIC just a few months after joining. (Big difference between promoting maker modules like Raspberry Pi and making big systems work.) Intel is pushing Edison and Curie, Samsung is pushing SmartThings, and critical mass remains an issue with too many other groups pushing IoT alternatives.
The OIC specifications perhaps weren’t the problem – the OIC membership definitely was. As long as there was perceived misalignment between the big mobile players, adoption would be in question. There was also the issue of comprehending the supply chain, from chips to software to infrastructure and big data expertise needed to make the IoT actually work. If you can’t get other players to give up their consortium efforts, the next best thing is to make a consortium-of-consortia.
Effectively, although not labeled that way, that is exactly what the new Open Connectivity Foundation is. They combined the existing OIC leadership in Intel and Samsung with the principals from AllSeen in Qualcomm and Microsoft and Electrolux plus some of the big hitters from the Industrial Internet Consortium in Cisco and GE Digital, and added ARRIS and CableLabs from the DOCSIS world. The big specification the OCF promotes is IoTivity, an open source software framework, and an alternative to AllJoyn and MQTT and other approaches.
Who’s missing? IBM is still lurking out there, and then there is the self-aligning Apple who wouldn’t be caught dead joining anything. I’d still like to see some more semiconductor firms like maybe an NXP or MediaTek. Asian device makers like Huawei and Xiaomi also come to mind. The shrewd player in this is Microsoft, who might help form a bridge between AllJoyn and IoTivity.
There was no technological announcement from the OCF, just new members and a complete domain redirect from the OIC. However, I do see merit in the consortium-of-consortia approach for the IoT given the telecom history I talked about – been there, have the t-shirt. It doesn’t solve every problem in the IoT or reduce the number of competing specifications, but fortunately it doesn’t increase them either. (Qualcomm was quick to straddle, saying this is not an AllSeen Alliance merger.)
The OCF has a consumer IoT slant, although if they make a dent it may carry the industrial IoT along for the ride. What the OCF does do is get many of the right people talking on what an end-to-end IoT solution will have to look like, and if nothing else happens, that is progress.