Just as Captain Kirk could always depend on his steadfast engineer, Scotty to manipulate the warp drives to deliver power to the shields of the starship Enterprise in evasive manoeuvre situations, so too will companies need to depend on their IT Operations team if they are to navigate their way through the Internet of Things.
The explosion in the number of applications for Internet of Things (IoT) within the B2B space means that sophisticated sensors can now be found in all kinds of assets from wind turbines, to jet engines. These sensors generate huge amounts of data, and in variable formats.
Take a Boeing 787 for example. These aircraft generate almost half a Terabyte of data every single flight. Multiply this by the number of flights and aircraft operated by an airline and that will give you some idea of the challenge at hand. This is the new frontier for IT operations.
Where to start? How about with an infrastructure that enables enterprise readiness? It is not enough to only look at the data lake, or the data warehouse in isolation. These days, it is generally acknowledged that what’s needed is an ecosystem, but that requires more management and input from an IT perspective.
If we are going to plug the data from Internet of Things into various analytical tools so that we can gain insights from it that can inform mission critical decisions, then it has to be available, reliable, authentic and correct. Of course, it has to make sense cost wise as well. And IT will need to find a way to securely transfer, process, store, and archive it.
Talking costs – open source on paper would seem a good fit cost wise. But will the IT Operations team be ready to manage the amount and frequency of updates of open source products, compared to enterprise grade software products? Specific update strategies would need to be put in place. With an open source stack, IT Operations need to be prepared to manage new and complex interfaces, and maybe even provide second and third level support themselves. These are the sorts of headaches that just wouldn’t come into the picture with enterprise grade vendors.
Then we have to consider that many IoT infrastructures will be operated as shared services within large companies. Large shared services and shared IT environments usually require multi-tenancy capabilities. Several business units or divisions will want to make use of the infrastructure, while restricting access to the data to particular groups of users. What sounds trivial might become very tricky in reality. Authentication, access management and so on could be challenging in complex analytical ecosystems or in a cloud-based environment.
Then we need to talk network. Network admins will have to be ready to make sure their infrastructure is able to ingest the data, transfer it into the repository and manage the peaks of traffic when users generate their Monday morning reports. This is where cloud-based solutions start making sense, for the sake of easing the strain on corporate networks. But wait – what does IT Security say?
IT Security will be looking at these new kinds of data and assessing whether the data needs to be kept under lock and key to protect the privacy of users and other confidentiality clauses. And while IT Security Officers have well established best practices for the handling of ERP, financial, HR data, that is simply not the case when it comes to sensor data. Why? Judging whether sensor data may or may not carry a significant insight on customers’ processes and even revenue streams (think output data of gas turbines), is far more nuanced.
In balance, while it is hard to argue that IoT will bring great (and positive) change to an organisation – the reality is that IT operations will not always love this brave new world. A recent study found that most companies are currently only getting benefit from a tiny fraction of the data that’s available from IoT.
There is a huge potential for uplift in competitiveness and productivity still to come from IoT – but companies will need to ensure that IT is involved in decisions made from the start before engaging their warp drives.
Michael Gerstlauer, Principal Consultant at Teradata
Image Credit: Ahmetov_Ruslan / Shutterstock