How can smart cities cope with the vast demands smartness makes on their IT resources? Partially replacing data centres with cloud computing is likely to be the answer, writes Barnaby Page
If data is the lifeblood of the smart city, the data centre is surely the beating heart that keeps it alive. It sits at the hub of all the urban networks, receiving data from (potentially) millions of sensors, devices, and other networks; analysing it; and issuing instructions and reports, not only to city management but to the city infrastructure itself.
The picture is, perhaps, an appealing organic one: it echoes the common metaphor of a city as a living thing. But the view is now widespread that it makes much more sense to take the smart city’s IT function partly out of the data centre and put it in the cloud.
The concept of cloud computing is simple and by now familiar: instead of owning, housing and managing IT resources such as applications and storage, the organisation (in this case perhaps a civic government) rents them from a service provider and accesses them over the Internet.
Appropriately enough for cities, it’s very much a utility model: just as we don’t need to understand reservoirs and pumps and pipes in order to be confident that water will flow out of the tap, an organisation using the cloud doesn’t have to worry which particular server in which remote data centre holds and processes their data. Hence the multitude of cloud jargon emphasising ’service’; the best known is ’software as a service’ (SaaS) and there are many variants such as PaaS (platform) and IaaS (infrastructure).
A move to the cloud has far-reaching implications, most notably improved agility, flexibility and efficiency once IT becomes an on-demand service, and the shifting of IT costs from capital expenditure to operating.
Also vitally important is the way that it frees the IT team from dealing with many of the nuts and bolts of upgrades and maintenance, which become the cloud provider’s responsibility; the in-house team can concentrate on more creative and strategic tasks.
Developers of applications that run in the cloud benefit in a similar way: for example, a software firm would not have to become deeply involved in IoT in order to employ IoT-generated data in the cloud: it could be confident that the data would be available in a usable way. The idea is that things in the cloud ’just work’.
But there are further aspects of the cloud that make it especially relevant to smart cities.
Maybe most important, there’s the sheer amount of data generated by all the city’s inputs; this has implications for storage requirements and for computational power, especially when real-time processing is required. Hosting this in the cloud – as opposed to a conventional data centre with IT resource owned and managed in-house – spares the city a massive rollout and a continuing operations burden.
The cloud is also inherently highly scalable and reconfigurable (in other words, you can easily add – or for that matter subtract – lots of resource quickly and easily, and it’s straightforward to change the way it’s structured). This suits the way most smart city projects are starting small and getting bigger, it accommodates urban growth, and it also allows for pilots, or plans that change.
Reliability is another key attraction of the cloud for smart cities. Systems running whole communities clearly need to be highly dependable, and a failed element can much more easily be replaced in the cloud than in a bricks-and-mortar data centre. For example, if a server goes down in the cloud, its tasks are seamlessly switched to another – and if a thousand go down, another thousand can take their place.
This kind of redundancy is viable because, in broad terms, all the cloud provider’s resources are available to all its customers, so they should rarely be sitting idle. In-house data centres can only afford a similar resilience to a very limited extent, because most of the time their backup systems would be doing nothing, an unproductive investment.
However, for cities – even more than for most users – blithely transferring the entire IT function to the cloud overnight is not a realistic option. The technology running the smart city is very often bolted on to much older systems which won’t sit comfortably with the modern, standardised, open systems approach of the cloud.
The solution recommended by most is compromise. As Margaret Ranken of the machine-to-machine (M2M) specialist Machina Research has written, in a white paper commissioned by Philips Lighting: “Making use of cloud infrastructure for new applications, while maintaining existing in-house infrastructure, at least in the short term, is the only practical solution.”
So, while she is adamant that the cloud has a role to play – “expanding an on-premise infrastructure to support the requirements of [a smart city’s] new applications would be a lengthy and expensive project carrying a significant risk of failure” – her prescription is the approach known as a hybrid cloud, mixing existing city IT infrastructure with new cloud elements.
This is only a halfway house, though. Most advocates of the cloud in smart cities are equally keen that the longer-term goal should involve moving much more IT out of the data centre and into the cloud.
In their paper “The Internet of Things Meets Cloud Computing in Smart Cities”, Schahram Dustdar and a team from the Vienna University of Technology argue that “current smart-city services are typically provided in single domains – for example, building management, transportation, and health care, among others. With such services, domain-specific application requirements drive all system-component design and determine most technical choices.
“The service-delivery process is rigidly orchestrated by domain solution providers. This model leads to many closed vertical systems. Scalability and extensibility are intrinsically limited in such systems, and the closed relationships between stakeholders stifle the creation of novel services.”
By contrast, they say, “service delivery for smart cities requires a new methodology aimed at delivering open and scalable smart-city services by encouraging collaboration between stakeholders in the Internet of Things and clouds. It exploits the prevalence of computing resources and software services distributed on the Web to break up the closed vertical service-delivery model.”
For example, according to Dustdar and his colleagues, building energy management (BEM) is an area crying out for the cloud philosophy. Implementing a separate system for each building or group involves complicated maintenance and makes sharing information difficult; putting a single system in the cloud, accessible to all, addresses that.
Another application demonstrating the cloud’s strength for smart cities is event organisation. Not only does this require a wide range of information (from maps to public transportation timetables to real-time data such as traffic and parking lot usage), which can easily be shared in a cloud environment, “the computing resources required are needed only during the events. Thus, the effort required to deliver such a service is justified only if the development and provisioning are efficient and cost-effective.”
How might all this work? Big names such as IBM and Microsoft (with CityNext) are already becoming involved in the cloud for cities. The European Union (through projects such as Storm Clouds and the Platform for Intelligent Cities, or EPIC) has taken a strong interest, too.
Then there’s the European-Japanese ClouT project, a consortium including companies such as ST-Microelectronics, Panasonic, and NTT, universities and cities, which aims to use the cloud as “an enabler to bridge the Internet of Things with Internet of People via Internet of Services”; it has deployed field trials in Santander, Genova, Mitaka and Fujisawa covering areas of interest to cities such as resource management, public safety and health.
And then there’s a host of specialist companies such as Romonet providing practical tools (in its case, software that helps data centre managers and cloud provider managers ensure their systems are efficient and cost-effective).
Much of the work done so far in supplementing the conventional data centre with a hybrid cloud to power the smart city – and eventually moving more out of the data centre into the cloud – is, of course, exploratory. (The Journal of Smart Cities has even devoted a special issue to examining the technical details.) But Dustdar and colleagues offer a vision of how it might work in practice.
Underlying it all are “domain-independent service-delivery platform providers, who present a new type of [cloud] offering that integrates IoT devices and infrastructures, processes data from a large amount of distributed data sources in real time, and lets applications employ both IoT and cloud resources on demand. The management of both IoT infrastructure and cloud resources is hidden from application providers.”
Then, utilising this infrastructure and resource, solution providers specialising in particular areas make “virtual verticals” – in effect software applications which might focus on areas such as traffic, parking, or energy. They “can reuse software services on the cloud and scale up services without investing in the computing infrastructure” – so not only will cities benefit from moving their own IT to the cloud, they will also benefit from their application providers offering software that runs there.
In the meantime, advises Machina’s Ranken, cities thinking about a move to the cloud should ask some fundamental questions:
Other important issues to consider before planning a cloud strategy and then selecting cloud providers include disaster recovery and resilience, security, and service level agreements (SLAs). The rewards are big, but the move is radical, and it’s vital to get the details right: we are, after all, performing surgery on the heart of the city.
Green heat from data centres
Environmental responsibility is a major issue for cloud providers, because their huge IT facilities consume so much power.
But so do individual cities’ data centres, and urban authorities retaining these on a large scale have an opportunity to turn a resource-hungry monster into a positive boon, according to Michael Beaven, co-leader of Arup Associates.
“By making data centres, and the sustainable generation to power them, part of the infrastructure of our smart cities, we could easily harness their excess heat and stop it going to waste,” Beaven suggests.
“Unlike many heat recovery sources, the heat from data centres is continuous and comparatively high grade. It can be easily upgraded so it’s hot enough to be useful in other buildings.”