Digitising the physical world: Forging a global M2M standard
By 2016 there will be an extra nine billion devices connected to the internet, many of which will talk to one another, rather than be used by people, according to projections by Cisco.
These 'machine-to-machine' (M2M) devices will be in cars, microwaves, pet collars, mining equipment, MRI scanners, tractors or any other place where it could be helpful to get a frequently updated stream of data.
If the last 10 years of technology development were about making it easier for people to exchange information with one another — Google, Facebook, Pinterest, Dropbox, and so on — then the next 10 years will be about making it easy for the physical world to transmit data to the internet.
The question facing telecommunications companies, networking infrastructure vendors and governments is how you manage these things effectively, and how you let them communicate so you don't run into problems like incompatibility or security foul-ups.
OneM2M to rule them all?
This year, seven standards bodies from across the world came together to try and solve this quandary by forming the 'OneM2M' organisation, whose goal is to hammer out the standards that will define how the internet's next few billion devices talk to one another without running into difficulties.
"M2M is already global by nature... what we develop needs to comply with requirements all across the world" — Luis Jorge Romero, OneM2M
"M2M is already global by nature... what we develop needs to comply with requirements all across the world," says Luis Jorge Romero, the steering committee chair of the OneM2M organisation. "If we do something regional, the basic fear is that each region by default develops their own, let us say flavour, and at the end of the day you will have difficulties in getting the right global standard. [This is] not something that is very efficient."
For this reason, OneM2M has brought together major standards bodies from Japan, China, Europe, Korea and the US to try and create a technical specification for M2M communication.
Companies from across the industry are participating in the scheme, including Alcatel-Lucent, Ericsson, HP, Juniper Networks, Motorola Mobility, Qualcomm, Samsung and Texas Instruments.
Each participating company is tied to a local standards body partner that provides information to ensure they don't develop technologies that "overlap with the work of OneM2M". If they have already duplicated OneM2M work, they will contribute it to the organisation, according to OneM2M documents (PDF).
By working with standards bodies across the world and with major technology companies, OneM2M hopes to create a standard that satisfies everyone.
Service layer architecture
OneM2M's proposed standards concern the 'Service Layer' architecture for M2M devices. Protocols and APIs will then be created that fit with this architecture.
Service layers are the systems used to pass M2M messages through a network, transfer data in and out of other IT infrastructure, present information to the administrator and communicate with other M2M clouds, according to Jon Howes, a senior consultant for Beecham Research.
Initially the Onem2M scheme hopes to hammer out standards for how individual M2M devices communicate their information, after which it will look at standardising how the M2M device authenticates itself with the network. If these projects go well, it will turn its attention to the mechanics of how the data is processed within the service layers —although there's some doubt about how far this part of the scheme may get.
Is the standard needed? After talking with analysts, the heads of standards bodies and technologists, many seem to think that basic standards for M2M communication are indeed necessary.
Already there are around 180 different ways of communicating, authenticating and securing data transfer between M2M devices and service layers, according to Marie-Paule Odini, chief technology officer of HP's communication media and solution organisation. Needing to support so many different technologies can put a huge cost burden on M2M technology companies, she said, describing the differentiation as "a showstopper" for broader uptake of machine-to-machine communication.
It should be possible to, for example, buy a fridge in China then move countries, taking take the fridge with you
"Standards that will simplify that and get alignment on the protocols on a network level and the wide-area network, and the data models and APIs, would [make M2M] something much easier to implement and support — it would reduce the cost," she says.
If OneM2M successfully creates a global standard for how these service layers function and exchange data with one another, then it should be possible to, for example, buy a fridge in China that communicates data with a local service layer, then move countries, taking take the fridge with you, and seamlessly plug into another service layer operated by another service provider or your own home network.
Although this example may seem trivial, it illustrates the interoperability benefits that standards provide: in the same way that any 3G phone can make calls across the world thanks to the universal 3GPP communication standard, future M2M devices should be able to use the OneM2M standard to exchange data with devices wherever they are.
However, forging a global standard is a difficult task, and there are already signs of tension.
Global versus de facto standards
For one thing, some M2M-focused companies are adopting proprietary technologies that are set to become de facto standards. The nightmare for OneM2M is that what happened with the cloud could also happen to M2M: most cloud companies are implementing APIs that are roughly equivalent to those pioneered by proprietary cloud operator Amazon via its Amazon Web Services (AWS) division.
Some M2M-focused companies are adopting proprietary technologies that are set to become de facto standards
"There really is no de facto standard in the M2M space," says Joe Rostock, chief technology officer of ATIS, the North American standards organisation and OneM2M-participant. "What's happening is there are proprietary solutions being developed."
Some of these standards are being built for specific industries, while others are being developed by companies to sell to a variety of business sectors.
"There really isn't one that's emerging as a leading standard," Rostock says.
Numerous technologies are currently potential de facto M2M communications standards, ranging from ZigBee's low-power mesh network, to Z-Wave's similar technology, to Neul's backing of the ex-analogue TV 'white space' part of the spectrum, to plain old 3G via companies like Wireless Logic.
Who's selling M2M?
Some companies are beginning to standardise on the service layer itself. In July Telefonica formed a global M2M alliance with other telecommunications operators that all plan to use Jasper Wireless technology for their M2M networks. Many of the world's leading M2M service providers — KPN, NTT DoCoMo, Rogers Wireless, SingTel, Telefonica, Telstra, Vimpelcom and Telefonica — are involved.
OneM2M's gambit is that Telefonica and the other ISPs will give their local standards bodies enough information about the technologies they are using for these capabilities to be implemented in the eventual world standard.
However, if M2M providers naturally converge around certain technologies, then a standard may not be entirely necessary.
"Off the top of my head... you don't need a standard because everybody is really building the solutions from a technical IP perspective," says Philip Cole, European sales and marketing director for Wireless Logic, a major player in the M2M space that uses SIM cards to get data in and out of devices. "I don't think a standard is necessary," he adds.
Others believe that OneM2M's plans don't go far enough and that more can — and should — be done in terms of creating standards for some of the more sophisticated technologies at play in this area.
Beyond the service layer
The feeling is that there are far more serious problems brought about by the rise of proprietary clouds for processing M2M data than there are in the basic communications layer.
"What I think is deficient about OneM2M is it doesn't form part of a bigger architecture for M2M that includes the applications and the data flows and the formats," says Jeremy Green, principal analyst within Ovum's industry communications and broadband division. "It's a bit telco-centric."
If you look at the list of companies involved in OneM2M, the majority are service providers or telecommunications companies and there are few participants from specific industries. However, this may be more representative of the overall shape of the M2M market than any real deficiency.
OneM2M believes that local standards bodies will be able to mesh their recommendations with the technologies being developed and sold by local M2M specialists. The gamble is that this can embrace many of the industries likely to use the technology.
"The interaction between sectors is something that has to be done," says Jon Howes, a senior consultant for Beecham Research. "If the standard is there for doing it in a common way throughout the industry, then that's a great thing for the industry."
Over the next few years OneM2M will develop the specifications and hope that they closely match the technologies being used. According to our interviewees, a standard seems to be the best way to help structure this market and avoid problems.
"Am I afraid that a specific vendor platform will drive the industry?" asks HP's Odini. "I don't think so. What will drive the industry will be standards like... OneM2M."
Read more on M2M: