The IoT Connectivity challenge
I have been working with Unify (now part of Atos), a company fully committed to an Open Standards strategy, for two decades. Without our commitment and investment into the most important communication standards, SIP (Session initiation protocol) and WebRTC (Real-Time communications for the Web), the vast majority of our current product portfolio (OpenScape solutions and Circuit), probably could not exist. With Unify having joined Atos, and with Atos being a key player in the IoT domain, it is obvious that I have started looking into what the situation is with regard to interop and standards in the IoT domain.
20 years back, with the introduction of IP technology in Enterprise Telco in the 1990es, international standards bodies were creating standards for Voice over IP (VoIP) in the second half of 1990es. I personally already participated in ITU (International Telecommunications Union, Geneva) in the specification of the first VoIP standards, best remembered as the ITU-T H.323/H.450 family. However, H.323 did not make the VoIP standards race. It was the IETF (Internet Engineering Task Force), that a few years later (1999) released its first version of SIP (Session Initiation Protocol). SIP was considered better suited for the Internet compared to H.323, although this was not only for technical reasons, but also a political discussion among stakeholders involved.
WebRTC is a new Standard (API) by the World Wide Web Consortium (W3C) for Real-Time Communication for Browser based applications, without requiring plug-ins. All major Browser vendors support WebRTC, such as Google`s Chrome, Mozilla`s Firefox, Microsoft`s Edge and Apple`s Safari. Unify with its cloud based Circuit Collaboration Tool leverages this new WebRTC standard.
However, SIP today still is the dominant Standard for Enterprise VoIP systems and we are being asked for SIP support in any larger customer RFP. But you should not believe that SIP based multi-vendor interoperability always works as “plug and play”, especially in case of more complex features and applications, such as Team Features or KeySets. This is due to lots of SIP options being available, based on which vendors are implementing their features (like from a Standards options tool box). As a consequence, there are still initiatives and forums, such as the SIP Forum with its “SIPconnect”, defining industry-standard SIP reference profiles for example for the SIP Trunking interface (interface between an Enterprise SIP Communication system and a SIP Telco Service Provider).
So why I am telling you all this ? In analogy to the Telco Real-Time Communications Standards domain, we also will need well accepted Standards for the IoT domain. If I compare the rather manageable amount of different functional entities that make up a SIP system (such as IP Phones, Soft Clients, Call Control Servers, MCUs, SBCs, Mobile Clients and probably a few more) with the almost endless list of different types of THINGS out there, which all have the potential to play a role in IoT communication (any kinds of Devices, Sensors, Cameras, Machines, Vehicles, Wearables, etc), you will understand where I am heading to. Standards, especially good and accepted standards, will be really helpful for all stakeholders, because they reduce fragmentation, complexity and risk of vendor-lock.
So what is the problem? The problem is NOT the lack of IoT Standards. In fact, I believe we likely have too many Standards Bodies, working on too many IoT standards. One reason may be that existing Standards bodies always will try to position themselves for new, promising technologies, such as being the case on IoT. The consequence is competition among Standards bodies, which is primarily good. So, on the pro side we can say that competition is a democratic approach and we also can say that the market will decide and sort it out. However, there is also a Cons side, in terms of that this rather large range of parallel standards bodies and standards activities may delay market introduction for some IoT vendors, due to uncertainty about which Standards to base on their IoT product developments. A further Con side of course is the lack of multi-vendor interoperability and thus the risk for customer vendor lock.
Based on a short assessment, I very quickly came up with a significant list of Standards Bodies, all dealing with IoT Standardization in some form. The list is for sure non-exhaustive:
- IETF (Internet Engineering Task Force), with its work revolving around its CoAP (Constrained Application Protocol), RFC7252 and extensions and enhancements
- W3C (World Wide Web Consortium), such as Web of Things WG and Devices & Sensors WG. Remark: I personally even expect, that W3C`s WebRTC standard will play a role also in IoT.
- IEEE (Institute of Electrical and Electronics Engineers), around P2413
- ETSI (European Telecommunications Standards Institute) M2M
- TIA (Telecommunications Industry Association)
- 3GPP with its push (supported by European Governments, such as Germany and UK), towards 5G also may play a significant role in IoT
Besides the well-known formal Standards Bodies above, a number of additional Industry Consortia were created specifically for IoT, working on a wide range of activities and deliverables, ranging from IoT Interface and API specifications to implementations of (Open Source) IoT SW development frameworks up to Certification programs and promotion activities. Example bodies are (but again, not limited to):
- Open Connectivity Foundation (OCF)
- LoRa Alliance
- Edgexfoundry by Linux Foundation
How do I see all this to continue? What are my recommendations or predictions?
- IoT Platforms System Architecture: Vendors should design their IoT product platform architectures in an interface and protocols agnostic way to be flexible to adapt to new standards when needed.
- IoT SW Frameworks: IoT SW frameworks should and will be leveraged, which for sure will currently not have support for all IoT (standards) protocols out there, but at least typically starting with a first set, adopting to new interfaces through Connectors, APIs, SDKs.
- Look at the segment and market you are targeting for. Some Standards Bodies or Consortia may be especially dedicated to the market or vertical you are interested in (e.g. WAN IoT/Smart Cities, Enterprise IoT, or thinking about geographic/regional centric views like NAM, Europe, etc. , although we actually should have left such regional views behind us and think in terms of global standards).
- Finally, and referring back to the Telco introduction of this article, the currently rather separated two domains, a) the IoT domain and b) the Telco / RTC domain will migrate. From the angle of the Telco / RTC domain, it will open up its architectures to integrate IoT traffic as yet another “media channel” to allow human information and even interaction. Similar, vice versa, from the angle of the IoT domain, some level of human intervention and human interaction will always be necessary also in the IoT data world. This migration will for example happen by means of intelligent IoT Edge Gateways/Servers/Bots via appropriate APIs, SDKs and Standards. I predict, that even RTC Standards such as W3C`s WebRTC will play a role in IoT, allowing real-time human intervention, interaction and collaboration based on real-time textual chat, audio and video, while IoT Data being carried within the WebRTC Data Channel.
Anyway, the long-term target has to be globally defined, implemented and well accepted IoT Standards, as a basis for world-wide IoT multi-vendor interoperability. This would help vendors to focus on Use Cases, business logic and customer requirements, rather than wasting time thinking about whether standard A, B or C is the right choice. Referring back to what happened in Telco domain, the market will sort it out, but it may be a long journey.