In amongst all the news about hyper-scale data centres and concentrations of data centres in key inter-connect metro areas, there has been an emerging story about the resurgence of interest in smaller local data centres. Initially this was about access, lower costs, local business Eco-systems, and good enough connectivity and latency. However the growth of the Internet of Things and its future potential has resulted in huge speculation about computing at the Edge and its impact on data centres.
Where traditional transactional applications, servicing widely dispersed end-users, have tended to favour centralised data centres with good interconnectivity, IoT is very different. The sheer amount of data generated (oil exploration and production companies can generate up to 250 TB of data a day) would put a tremendous strain on existing networks by itself. The fact that much of this data is both transitory in nature, and often required by only a small number of end users or other sensors, makes its transmission to and storage in huge centralised data centres highly inefficient.
Suddenly local data centres become an attractive location for IoT data and transactions. More than that, existing Point of Presence facilities (PoP) could become effective Edge data centres with a certain level of new investment. So far, so what? This is not a new or original idea. Analysts and commentators have expended many column inches on the topic. We are seeing local data centres begin to change their messaging. There are specialist Edge data centre companies like EdgeConneX emerging. However the potential bonanza for local data centres might not always be realised.
We need to look back to the emergence and development of machine to machine (M2M) technology to understand the pitfalls for unwary data centre operators looking to take advantage of the potential offered by IoT. M2M is fundamentally a sense and respond technology that enables engineers, production managers, and safety staff amongst others to optimise industrial processes. IoT has broadened the scope, scale and flexibility of the old M2M model, bringing in a range of sensing devices unimagined even 5 years ago. It has also radically expanded the requirement for analytics. But fundamentally it will still be the engineers and operations managers who will be controlling and making the decisions about when, where and how to deploy IoT infrastructure. By and large these are not the people who have traditionally been involved in decisions about information systems.
This means that data centre operators, hosters and MSPs will need to identify, understand and communicate with a different group of decision makers and influencers from the ones they are used to. This is not a simple case of changing the names on the target audience list and repurposing existing content and propositions by changing the taxonomy. Technical and business issues will be different. Requirements and expectations will change. Investment hurdles and decision making process could also differ significantly from those you see in existing Cloud and Internet based transactional systems.
To misquote John F. Kennedy, "don't ask what the IoT can do for my data centre. Ask, what can my data centre do for the IoT." Take time to review your go-to-market strategies and value propositions. Make sure you have framed them from the customer's perspective. They could well be quite different from the ones you have been used to up to now.
This post first appeared on the old Cassini Reviews website.