It’s fair to say IT doesn’t have many facets people would call “sexy” but there’s always the current buzzwords floating about the boardroom. The CIO / Technology Leader role naturally needs to understand these trends and ensure the board are fully briefed on any opportunities or risks that may arise from new and disruptive technologies. I’ve truly lost count of the times I’ve needed to explain blockchain to CxO level leaders (worryingly, some of these leaders were in tech leadership roles!) There’s one area of technology that I think is difficult to approach for funding or get people emotive about, but for me, it’s truly the engine room of any large enterprise; Integration Hub.
Take any large enterprise system and start the steps towards replacement and it’ll send a shiver down the spine of any IT project manager, leader or specialist. How many refresh or replacement projects have you seen go over budget, miss deadlines or even fail completely? Here’s where a well abstracted integration hub comes into its own.
The simplest analogy for good abstraction comes from the concept of sending something through the mail. You, as a sender, need to adhere to certain rules – place it in a suitable package, ensure the address is present, has a postal code and pay whatever postage is necessary. As long as you adhere to these rules, the postal service will be able to route and deliver the mail. “Leaky abstraction” is the enemy here (arguably, the postal / zip code is an example of leaky abstraction but it’s a simple analogy most folks understand the concept of). Imagine if the postal service needed you to write the routing information on the package. If each and every time the package needed to pass through a sorting hub or delivery centre, you were expected to ensure that information was placed on the package. OK, a little more work I hear you say but you could probably handle it. But say you send 100 parcels a week to a remote location so, for speed, you pre-printed a load of labels. Then the postal service changed their routing. This is where the leak becomes damaging – because decisions about what should happen within the system have been made from the outside, the core of the system has effectively lost control. It will perform poorly through no real fault of its own.
Leaky abstraction is the cause of so many headaches in enterprise software. In my role, we see this most with the integrations into logistics providers. Retail teams should care very little with the pick walk the warehouse operative is taking, or unless there is a specific customer promise, which courier is handling the final mile delivery. The logistics platform should not be handing data over to other upstream systems (we like to call this “bunny hopping”). Core data should be sourced from a single source and fed to all downstream systems equally – making each component in the ever-growing web of enterprise systems replaceable without affecting others as long as the new system can replicate the data in and out of the old system via the integration hub. And yes, it is even possible to take this approach with your ERP. It takes some discipline. It means keeping a tight control on technical debt that breaks this model. We all have tactical deployments going on that need to overcome a short-term hurdle but make sure you go back and plug that leak when you get the chance. Your users will usually thank you for it as the tactical solutions are rarely the most user friendly anyway.
So next time your CEO is asking if blockchain is going to revolutionise the way you work (hint: unless you work in massively distributed currency exchanges with a low transaction throughput, it really won’t) maybe steer them towards a different technology. “Let’s talk abstraction, baby!”
Comments
Post a Comment