Software-as-a-Knot
--
“We forge the chains we wear in life”, Charles Dickens, The Cricket on the Hearth
Why do established organisations find it hard to change ? Why do so many digital transformations programmes fail ? The consultants have an answer for this — it’s the culture. But it really isn’t, and surely you must have learned by now not to believe a consultant who asks you to trust them on something that nebulous and hard to measure ? No ? Then maybe you need to come back and read the rest of this blog once you‘ve reached the anger stage.
Sure, there may be a few people stuck in middle management jobs who are doing their best to hold everyone else back, but that’s not a culture. We can only blame culture if what we mean by that is the atrophied amalgam of human capital, software systems and infrastructure that sits at the heart of most enterprises — otherwise reasonable people just accepting how bad their systems are and spending most of their time on workarounds and make-work — a kind of Cyber Stockholm Syndrome. To understand what’s really happening, we have to look at how enterprise IT systems are built. Sometimes they’re done well — the architects really manage to deliver something that supports what the business needs. Sometimes, they’re not so good and you have to adapt the business to meet the software. After a few years of this though the workarounds become the norm and no-one can remember what good looks like anyway. Then there are the monoliths. In a stroke of genius in the late nineties, the consultants (them again) realised CEOs didn’t trust their CIOs to deliver. “Golf course IT strategy” was born — bypassing the IT department altogether and just buying a monolith that does it all. Sure, you need to adapt your processes to whatever the monolith tells you[1], but at least it works and you know exactly what it costs…you’re sure that’s what they told you in the 19th, but it’s all a bit hazy.
So what’s this got to do with digital transformation ? Or any kind of business change for that matter ? Well, it doesn’t matter whether you have off-the-shelf, custom-built or monolithic IT — after a while your business processes and your IT end up deeply intertwined. If you want to change your business, you’ve got to change your tech. Changing a monolith (let’s call this “software-as-a-blob”) isn’t going to be easy. The vendor is usually sitting on decades worth of source code, and has other customers to think about. If they’re not all asking for the same features as you, good luck with that change request. Bet you wish you’d gone with the in-house build now eh ? Those ponytails in the IT department are suddenly starting to look like a group of wise elder statesmen. Well, er no. The truth is rather worse. All those in-house systems (or bought-in applications) need to talk to each other. That means you need to build interfaces between them - or if you’re really lazy you can just copy the data structures from one app to another. Either way, if you want to change one of those systems it’s going to have a huge knock-on effect on dozens of others. Your architecture has been made brittle by all those fixed interfaces that create branching fault lines through the whole thing. Even the smallest of changes to a single application can take years and cost millions. That’s why well established enterprises can’t change. One organisation we spoke to spends $1Bn a year on integration. That’s not their IT budget, that’s just for maintaining the interfaces between systems. We call this “software-as-a-knot”.
It’s not just a business change thing either. If those systems that you can’t change rely on older, insecure infrastructure that you can’t replace or patch, you become vulnerable to malware intrusion. Sometimes you get a blob surrounded by lots of knots where even the blob vendor can’t patch your installation because it would break too many interfaces — again leaving you vulnerable to attack. The organisations that make the headlines when they get pwned are often unpatched because patching might bring down their entire IT stack.
The idea that there is some Gordian knot at the heart of your business that’s holding you back may come as a surprise to some business leaders. What’s more worrying is that it also surprises a number of senior IT leaders. They’ve spent the last couple of decades preening over their rather neat-looking service-oriented architecture — usually an enterprise service bus with lots of systems connected to it. This just skims the surface though. While it’s true there are physical connections between the systems and the bus, the data picture is rather different. People who don’t work with data for a living (and this includes most IT folk who are either infrastructure or software specialists) think that data flows like water. It doesn’t. It spreads and replicates like a virus — copies here, copies there, forks, mappings, derivations, checksums, etc. Because everyone was fooling themselves they had a nice neat architecture, the knot was allowed to grow unchecked for decades. At one point, a major logistics organisation I knew of was maintaining twice as many interfaces as it had applications. They actually wanted to switch to a monolith, but couldn’t.
There’s also a cost and trust problem. CTOs don’t get many chances to tell the board that spending $XXm on a new platform will solve all their problems. Especially given all those previous $XXm purchases were supposed to do that. The IT department has to resort to under-the-radar operational expenditure to fix the problem, and that just ends up adding more knots and tightening the knots that were already there. Even when someone is prepared to stump up the cash to build something new, the old systems seem to linger — as a result of missing something crucial for compliance in the cutover more often than not.
So, I haven’t ranted on about the state of enterprise IT for no reason. This has been a problem that the Telicent founders have been battling with in various companies and government departments for decades. We founded Telicent in 2020 to tackle the problem of blobs and knots. We’ve put openness to the fore. Open data standards, and open source software. Our core platform connects to legacy systems and makes their data available though standardised APIs and smart caches. The platform allows you to behave like a digital challenger, operating in a modern, greenfield tech environment, but with access to all your valuable data. It allows you to work in parallel with your legacy, replacing functionality step by step as you gradually phaseout the knots and blobs — so no need for expensive capital outlays on “big bang” programmes. We’re just setting out on this journey and we’re looking for companies who want to pilot new technology and help us shape the platform.
Ian Bailey, CTO, Telicent
[1] “I mean, the joke about SAP has always been, it’s making 50s German manufacturing methodology, implemented in 1960s software technology, delivered to 1970-style manufacturing organizations, like it’s really — yeah, the incumbency — they are still the lingering hangover from the dot-com crash.”
