AI
Image credit: Stock

David Socha discusses the proliferation of technology platforms across every aspect of IT, OT, data and analytics and considers the lack of a consistent definition of what exactly a platform might be.

This article was originally published in serial format on smart-energy.com and in Smart Energy International 3-2018

Everything’s a platform nowadays, huh? There are smart grid platforms; smart metering platforms; all-encompassing utility IT platforms; big data platforms; analytics platforms; IoT platforms of course; AI platforms…obviously. In fact, I could probably fill my allocated word count for this column and still not name every variation. But what is a platform? Or perhaps more importantly, what do vendors mean when they talk about platforms? Because that might not be the same thing.

The Truth is Out There

Let’s see if we can make some sense of this whole platform concept and at least have a go at cutting through the marketing doublespeak. And let’s start with trying to define what a (technology) platform actually might be. Techopedia’s definition is a good place to start. It says a platform is “a group of technologies that are used as a base upon which other applications, processes or technologies are developed.”

I can work with that. Take, for example, a smart metering platform. The above definition might suggest such a thing to be a group of technologies that do some undefined number of basic tasks that are likely to be common to a lot of smart metering processes. It might not do the more complicated stuff, or stuff that’s specific to your business’s or your regulators’ unique requirements – that would need to be developed on the platform. Or perhaps we don’t need development. Perhaps some of that more complicated stuff might come as optional additional pre-configured apps or modules from the vendor. With me so far?

Brand X

Well yes David, I hear you say – but what are those common basic tasks? Who gets to define what functionality is basic (i.e. part of the platform) and what isn’t (i.e. stuff I have to buy or develop separately)? And how do I know if the platform is going to cover all the things I need it to cover? Who gets to decide the minimum set of functionalities allowed before a vendor can call their product a smart metering platform at all?

Is there some International platform police organisation that I don’t know about that makes sure everyone is conforming to the same set of standards?

Hopefully, you’re starting to see the problem. In the absence of said platform police, there is no way to immediately know what a vendor means – or what they’re avoiding telling you – when they proudly introduce their particular Brand X platform to you. Some platforms are comprehensive, proven, flexible, scalable and extremely valuable. My own employer’s analytics platform is one of those. But then I would say that, wouldn’t I? Many more platforms are a rough cobbling together of some tools and applications that the vendor has created, procured, inherited in mergers and/or is still working out how to integrate. And they may or may not cover all of your needs.

Deceive, inveigle, obfuscate

Here’s where we get to that part about cutting through the marketing doublespeak. It’s absolutely critical to go into a conversation with a platform vendor with eyes wide open. Some platforms are based on data integration and flexible, managed access to that data, creating a…eh…sound platform for all sorts of applications and processes to leverage a single view of an enterprise’s “truth”. Others focus on applications, development toolkits and point-to-point data integration, or can only integrate specific types of data, limiting your view of the business by maintaining (or even creating new) silos of information. Some are tied to vendor-specific ways of developing new front-end apps, limiting your ability to use existing tools your business might already have paid for – and that users know and love and deliver valuable insights with. But each will be called a platform, irrespective of what capabilities they have and what functionality they offer.

I want to believe

So how to navigate this ambiguous world of confusion and obfuscation? We all want to believe that the right platform for our needs exists out there somewhere. And also, that we can evaluate apparently comparable/competing platforms objectively to come to the right conclusion for our business.

So we still haven’t got as far as coming to any real conclusion on the definition of a technology platform, simply because… well…it doesn’t matter what we decide, does it? Vendors have always called whatever they sell whatever they think will make you buy it. And everything’s a platform nowadays, huh?

Take this example: While I still subscribe to the definition of a platform we used earlier in this article:

“a group of technologies that are used as a base upon which other applications, processes or technologies are developed.”

… others are determined to convince us that a platform has nothing to do with technology at all and is, in fact, a business model.

And hey, who is to say they’re wrong?

So how to navigate this ambiguous world of confusion and obfuscation? First, let’s assume we are indeed still focusing only on technology platforms. Now, you may not be the IT expert in your business. And if that’s the case, you shouldn’t need to be either.

I get that. But, so does the vendor. And at times they can…eh…gloss over a few things that it might have been more useful for you to know before you do engage with your IT colleagues and ask them to help you deliver some new value for your business.

With that in mind, let me propose three areas for consideration and questions. There’s more to it than this, of course. But I hope these three can help you to cut through some of the pretty marketing images and overhyped buzzwords to understand what it really is that your platform vendors are offering you – and what they’re not.

Coverage

At the simplest level, ask your vendor which key groups or layers of components their platform provides. Is it just a suite of useful applications and software development kits (SDKs) so you can make more applications? If so, I’d suggest that what you have there isn’t a platform at all. It’s a suite of applications. You’d be surprised how many vendors will fall at this hurdle.

Of course, a suite of useful applications may well be attractive to you. And that’s fine. But let’s consider what it is that actually makes those applications useful. It’s data. Does the platform also include comprehensive data acquisition and processing tools and capabilities? In other words, will the platform ensure that the right data are sourced, processed and securely delivered to its applications at the right time and in the format they require? Or do you need to do that yourself?

Finally, on this topic, does your chosen platform include appropriate data storage and curation? Many don’t. Many will expect you to already have your data in an ecosystem that allows fast access to already curated data. Without it, their platform may still be able to function, but data may be sourced from point-to-point integrations: increasing costs; multiplying the likelihood of duplication and errors; and reducing performance. Or, there will be an expensive change request on the project to deliver said data ecosystem. This is perhaps the single most common ‘misunderstanding’ encountered when investing in a technology platform.

Openness, connectivity and interoperability

Now let’s think about the fact that you might not only want to use the applications that come in the box, or that you can develop with that neat SDK. Does your vendor’s platform allow your existing applications to take advantage of the platform’s infrastructure too?

Can your Data Scientists apply Open Source analytics tools to the data in the platform, to complement those the vendor may have provided? And what of cross-functional analytics? Let’s illustrate this issue with the familiar example of smart meters, as I have before. Can you quickly, easily and regularly extract smart meter data from your platform to create new insights across a wider environment including, say, billing and geospatial data from other sources? Or is the data in your new platform now stuck there forever, next-to-impossible to extract without specialist skills?

Real scalability

Do you know the data volumes your platform needs to be able to handle? Perhaps more importantly, do you know how much it may have to handle in the near future, as digitalization and the IoT deliver unprecedented amounts of data to the enterprise? Again, it may well not be your role to know such things. And that’s OK. However, if you’re the business user assessing the capabilities of a new platform, make sure you – or a colleague better placed to ask the difficult questions – do ask searching questions about scalability.

Any platform vendor should be able to demonstrate lightning-fast responses on small, or even relatively large test data sets. But, sticking with the smart metering theme: Enedis is rolling out 35 million smart meters across France. TEPCO is rolling out 27 million across the Tokyo area. That’s a lot of data. IoT platforms may have to be able to handle even more, as more and more sensors send back their data to the enterprise. And what of those cross-functional analytics platforms, potentially looking across all of that data? Scalability, scalability, scalability.

Over to you

So, should you find yourself talking with a platform vendor any time soon and it seems they might have something that could be of use to your business, I hope you’ll consider the three lines of questioning I’ve outlined. First, in a simple three-layer model, does the platform deliver at the application, data processing and the data storage & curation levels? Second, is it open, accessible and interoperable, or a closed shop? Third, does it scale? I mean really scale?

If you can get the answers you need to these questions, it’s a fair bet that you’re on the route to securing a platform that could deliver what you thought it might when you read that first marketing headline or heard the first elevator pitch. If you can’t … well, don’t say I didn’t warn you!

ABOUT THE AUTHOR

David Socha is Teradata’s Practice Partner for the Industrial Internet of Things (IoT). He began his career as a hands-on electrical distribution engineer, keeping the lights on in Central Scotland, before becoming a part of ScottishPower’s electricity retail deregulation programme in the late 1990s. After a period in IT management and consulting roles, David joined Teradata to found their International Utilities practice, later also taking on responsibilities in Smart Cities and the wider Industrial IoT sector.