Plugging M2M sensors into the network
There is no shortage of innovative uses of M2M technology, although you would be forgiven if you chose to replace ‘innovative’ with ‘gimmicky’. One of the most recent applications in the news is the Egg Minder – a $14 egg box that contains tiny sensors at the bottom of each egg-shaped depression and which will notify your smartphone when you’re running low. Yes, it’s fun and silly, and of no use whatsoever to the commercial sector.
Except that the Egg Minder is being developed in partnership with GE, who is investing a considerable amount of money and effort in promoting ‘the industrial internet’ –GE’s own take on the convergence of Big Data with M2M and the so-called ‘Internet of Things’.
Behind all the tabloid-friendly stories of consumer gadgets like the Egg Minder or the Intelligent Fridge, M2M is gearing up to become a sizeable and significant disruptor of services and businesses. Telecoms infrastructure vendor Ericsson grabbed the headlines with its forecast of 50 billion connected devices by 2020. A significant part of this will be M2M modules, as connected devices move from the consumer sector out to networked industries and eventually the ultimate dream of a networked society.
Mobile operators are keen to position themselves as facilitators of M2M services, after all, they are in the business of selling capacity on their licensed networks. In June, Deutsche Telekom launched a range of M2M development kits for programmers looking to develop cloud-based M2M applications. The kits consist of an Arduino or a Cinterion board with a GSM chip, a SIM card, and access to its M2M Developer Platform.
The launch coincided with the publication of a report from market research firm Infonetics that pegged the global mobile (cellular) M2M module market at $1.5 billion at the end of 2012 – an increase of 25 per cent from the previous year. This growth is coming mainly from the automotive, transport and logistics sectors in Europe and North America. Infonetics projects a cumulative $2.6 billion will be spent over the next 5 years on mobile M2M modules in these verticals, although the increasing focus on Smart Cities is enabling opportunities for other verticals, such as the utility sector.
Today’s mobile M2M modules operate predominantly over old 2G cellular networks, as the amount of transmitted data for most applications is very low. This allows for cheap data and cheap hardware. However, as we start to see cellular operators phase out their 2G networks and re-farm the spectrum for 4G services, M2M is moving to 3G and beyond. By 2017, over half of all mobile-based modules are expected to run on 3G networks, but many vendors are already preparing for more widescale 4G (LTE) deployment and consequently 4G M2M modules are seeing the fast grow rate.
Yet cellular-based mobile M2M actually only accounts for about 15 per cent of active M2M connections today. M2M also works perfectly well over unlicensed frequencies using wi-fi, Bluetooth or Zigbee, and private mesh networks. Then there’s the new Weightless wide-area wireless networking technology designed specifically for M2M, based on a modern set of wireless air interface technologies that its proponents say provides a better future option than 3G or LTE.
The migration of smartphone-heavy mobile traffic to 4G will prove much more spectrum efficient than 2G or 3G for telecoms operators, and hopefully result in more spectrum for M2M – but it’s still competing for a finite resource. The switch to digital TV in many countries has made available more spectrum resources for the regulators to auction off, but it has also given rise to another interesting possibility: white spaces spectrum.
This is the name given to spectrum that exists between the frequency bands used by broadcasters, and which is now seen by a number of countries as providing an alternative route to mobile connectivity. The spectrum is unlicensed and available to all, with a sophisticated database managing all possible interference and co-location concerns. Trials are already well underway to use this spectrum not only for providing broadband access to rural locations (its low frequency means excellent propagation characteristics), but also as the backbone for M2M connectivity. It’s also relatively technology agnostic, with successful trials using wi-fi variants, WiMAX and Weightless protocols.
It’s also important to note that M2M even works over fixed-line networks, dispelling the myth that it is a mobile-only solution. The likely scenario for utilities will be a blend of different mobile and fixed-line technologies operating over a range of frequencies.
There are though, strengths and weaknesses for each of the different underlying network technologies. The right choice very much depends on the specific application and user case. Security and performance reliability are just as important as cost effectiveness.
The next major point is what are these M2M sensors built on? What operating system underpins them? What transport protocols are being used? The argument that’s currently raging is over XMPP and MQTT – two separate standards that can provide the interoperability needed for widescale M2M deployment.
Extensible Messaging and Presence Protocol (XMPP) is a messaging protocol based on XML and which was originally called Jabber, developed by the Jabber open-source community for instant messaging and chat applications. The commercial arm of Jabber was acquired by Cisco in 2008. Google Talk was launched on Jabber, but this year Google dropped it as a supported protocol.
But the supporters of XMPP want to move it away from chat and use its specific characteristics to support M2M and the Internet of Things. Being extensible, there is now support within XMPP for sensor data, optimised for reading meters. In fact, anywhere there is a ‘demand-respond’ situation, XMPP could be an ideal choice. It allows for precise identification and gives a high level of data security.
However, there’s also Message Queue Telemetry Transport (MQTT), another open messaging protocol. Where it differs from XMPP is that it has been specifically designed for M2M applications. It is optimised the transfer of data from sensors and embedded devices across high latency or congested networks. There is a degree of disquiet on the web over the development of MQTT, as it was created by IBM and took a long time before it was developed into a open source version.
Perhaps the good news is that there is unlikely to be a winner, and that the protocols should interoperate. XMPP and MQTT also face competition from the likes of CoAP and OpenSCADA – no one solution looks like dominating.
Added to the complexity surrounding the mobile air interface and transport protocols are issues such as data formats and structures. As Paul Duffy, Senior Technical Leader for Connected Energy at Cisco, says: “Connecting sensors and objects opens up an entirely new world of possible use cases – and it’s precisely those use cases that will determine when to use the right protocols for the right applications.”
Bottom line, there is no one size fits all solution.
The utility sector has all the right components to be a key target for M2M vendors. For a new technology solution to work, it has to address existing problems. Utilities provide mission critical services that cannot fail, yet rely on complex and at times aging infrastructures that need constant monitoring. Having to rely on people to perform all this monitoring is expensive and time-consuming; far better to have remote sensors embedded throughout the network that will operate 24/7, feeding data back to an operations centre.
Then there’s the prospect of integrating renewable energy sources into legacy networks. Renewables are often sited at remote locations and operate intermittently, making monitoring and control difficult. By placing sensors on the right equipment, utilities can be better prepared and make more accurate use of predictive modelling for improved operational efficiency.
M2M is a key component for both smart grids and smart meters. True, smart meters have proven pretty useless without corresponding ‘smart tariffs’ – and the necessary changes to the regulation that go with it. After all, there needs to be a financial benefit to the customer. But they will eventually start to make a difference. Together, these elements will help create Smart Utilities, and mobile operators are keen to get involved – for obvious financial reasons.
But choosing a communications network partner for M2M could also depend on the business strategy of the utilities, specifically when it comes to capital expenditure versus operating expense. It may be financially beneficial for a utility to build out its own network, using unlicensed spectrum technologies like wi-fi or fixed line connectivity, rather than enter into an ongoing service contract.
However, there is far more to M2M than automated metering. Smart Metering can handle multiple functions at once: improvements in billing and service management, device deployment into more secure locations, asset control, and in-field service diagnostics and reliability.
The promised widescale adoption of hybrid-electric vehicles will also put pressures on power grids and could provide an additional catalyst for the roll-out of smart meters, as customers will expect to be able to recharge their vehicles at home and power providers will have to design new payment models to handle this additional requirement. Not to mention concerns over grid meltdown if more than a handful of Electric Vehicles are on a quick charge on one street.
Then there’s home automation. Smart utilities can use M2M sensors to provide support and control for HVAC (heating, ventilation and air conditioning) services. They face strong competition from telecoms operators though, with the likes of AT&T (Digital Life) and Orange (My Plug) already building their new business units and signing up customers. The ‘connected home’ is becoming a key battleground in the fight for new M2M markets.
With utilities, telcos and ISPs all enjoying valuable access to domestic buildings through their existing connections, and equally importantly having existing billing relationships with homeowners, it makes sense for them to want to leverage their assets into home automation. Once you have control of the central home hub unit and become the de facto billing point, then it’s just a case of plugging in additional devices, sensors and applications. The telcos have already discovered that consumers love simplified billing – first through the bundle of phone, internet and TV into ‘triple play’ products (and then the addition of mobile to form ‘quad play’), and recently the idea of family shared tariffs for mobile accounts has started to gain traction. Utilities need to be very aware of the dangerous lead that telcos enjoy here.
The future is all about connected devices – and lots of them. And once we have our Internet of Things, we need to be able to be able to create viable business models around tariffs and fees, to provide actual value to the users of these devices. Whilst technologies will continue to develop, so to will industry disruption. If you’re not already engaged in the M2M sector, you are at risk – and not necessarily from one of your traditional competitors.
As Miguel Blockstrand, Head of Product Line Device Connections at Ericsson, and the sponsor of the annual M2M Challenge competition, says: “Only our imagination sets the limits for what can be accomplished in M2M.”