By Monika Ermert, eLance Journalist – Eight working groups of the Internet Engineering Task Force (IETF) are covering aspects of the Internet of things (IoT), full time or partly. And as if that wasn’t enough, a ninth one will be addressing issues in Low-Power Wide-Area Networks (LPWAN).
Despite having worked on IoT issues since 2005, the IETF could not “rest on its laurels” earned from developing and adapting the I in IoT, said Carsten Bormann, IoT expert from the University of Bremen and Chair of the Things to Things (T2T) research group of the IETF's research sister organization. Here is a little overview over the gaps to be filled and the vertical integration trend to be pushed back against.
IoT is different from other Internet technology areas in that it is not as uniform, and not made for a single purpose as the good old Internet, explains IETF Chair Jari Arkko. Well-known for having talked to his washing machine for years and being interested in not leaving the communicative in- and output of his own car to car companies alone, the Ericsson engineer underlined the multifaceted network of things.
Up and down the stack
Contrary to the engineering task for the network of networks, IoT is affecting multiple layers of the protocol stack from the radio to IP layer, further up to the transport and middleware layers. Improvements on radio and cellular come from 3GPP, just one example of another speciality for IoT: standards bodies have to cooperate. Links that have been made include the Web of Things initiative of the World Wide Web Consortium (W3C).
In addition, the systems belonging to IoT vary widely, explained Arkko, from heavy-duty industrial machinery to gadgets and even clothing. “The IETF's role is more on the fundamentals than the applications or radio arrangements.” But to a large extent, the world has understood in the meantime that the Internet was the connecting glue between various IoT systems, Arkko said.
A decade of standards work style
The IETF was early in standardizing for the network of things, underlined Bormann. So far, the IETF has focused on “filling gaps to make IP viable on IoT environments, defining fundamental protocols for application-layer interoperability between and with things, creating the web of things and defining (and refining) robust, usable security mechanisms, which are required to enable a real, scalable IoT”.
Bormann, together with his T2T research group co-chair, Ari Keränen, recently provided an overview of the WGs that have been pushing the IoT work ahead in the IETF. Of the eight WGs, the ones making IP viable for IoT include Roll (for the routing level) and 6LoWPAN (adapting IPv6 to IEEE 802.15.4) and its follow-up 6Lo (adapting also to Bluetooth, ITU-T G.9950/Z-Wave, etc.), as well as 6Tisch (IPv6 over IEEE 802.15.4.e).
The space and the attention given to it is so large that a special IETF WG, LWIG, is documenting guidance information relevant to these standards.
The newest working group LPWAN, is an effort to find a common abstraction based on IPs for the various connection technologies – NB-IOT, 802.15.4 (Wi-Sun) LoRa and SIGFOX. There was a desire for common services and abstractions for the upper layers, argued participants of the upcoming new WG. According to Bormann, application layer interoperability issues are currently dealt mainly within in the CoRE (Constrained Restful Environments) WG.
Security and other gaps to fill
Security is also high priority in the IETF's IoT work, underlined Bormann. IoT security is on the agenda of at least three WGs: DICE (DTLS in Constrained Environments), ACE (Authentication and Authorization for Constrained environments) and COSE (CBOR Object Signing and Encryption). The latter adapts object signing and encryption methods used for Java objects to the IoT, using the slighter Concise Binary Object Representation (CBOR). IoT systems can also benefit from more light-weight encryption algorithms, namely the elliptic curve crypto algorithms, in the Crypto Research Forum (CFRG) and other work on security.
But there are also additional optimizations to be done in the security space, acknowledged Hannes Tschofenig, IAB member and engineer at ARM. Tschofenig is the author of a new document which not only explains how to protect CoAP messages using DTLS 1.2, but also “defines a profile of DTLS 1.2 and TLS 1.2 that offers communication security services for IoT applications and is reasonably implementable on many constrained devices.” Only the appropriate extensions need to be used, Tschofenig added.
However, one can always do better security-wise. Including better security remains one of the crucial areas, according to IETF Chair Arkko. Asked what he thinks are the challenges of IoT standardization, he points once more to interoperability as a core task for the IETF. Practically solved at the lower stack, it was not yet possible to integrate larger systems and “build applications that could leverage many individual device types”.
As Arkko stated in his report for a workshop of the Internet Architecture Board on “IOT Semantic Interoperability”, data formats and object models too often differed in ways that made it difficult to integrate individual pieces into larger systems. Looking for commercial benefits, vendors tried to live off their special “walled” solutions. Therefore, the main task for the IETF could be to lower the barriers of entry and allow more players to hook-up to the new network of things – just as it did for the network of networks.
