On The Edge Of Things | Blog

Why Legacy MDM Will Become Extinct

Written by Steve | Sep 21, 2022 1:55:40 PM

I always assumed Homo Sapiens were the predecessor to the Neanderthal, but in Harari's book Sapiens he tells the story about how the Neanderthal and Sapiens actually lived side by side. Neanderthals were stronger, faster, better climbers and by all accounts, physically superior. But there was one thing Sapiens had that the Neanderthals lacked: the capacity to coordinate by the thousands at the edge of battle. Limited in capacity to coordinate in large numbers, the Neanderthals never had a chance. They were wiped out by our Sapien ancestors and as they say, the rest is history.

Something similar is happening in many technology segments in general and Mobile Device Management (MDM) in particular. If you're running a legacy MDM, you may have noticed already. Whether it's overly-complicated interfaces inspired by eBay and AOL, updates that are disruptive to operations, or innovation that happens at the pace of an underwater Tai Chi class, legacy MDM is crumbling at scale and losing ground, unable to catch up due to the weight of its own construction. The architecture doesn’t carry the capacity to efficiently scale, let alone run autonomously at the edge, which is what most businesses will need in the future to survive and thrive.

It's not that the designers weren't smart or talented or visionary, legacy solutions simply had the misfortune of being conceived at the wrong time. Back then few dreamed of an internet of things driven by autonomous, learning machines. Fast forward to today and MDM is staring down a serious inflection point, and if you're not moving towards edge oriented technologies and autonomy, your MDM is probably a dead man walking.

Using an edge-first approach to design and development--one that minimizes central server dependencies and a constant connection to function--is faster, and cheaper to run. Not only that, it’s more responsive, private, secure, and resilient than centralized, single-server dependent models deployed in legacy systems. It's what the FAANGs and the other tech giants deploy and it's cascading down to us, faster than you might expect.

Edge oriented technologies are faster because they push the data and logic to the place where the data is created, at the point of interaction. Increasing device autonomy radically decreases all costs associated with latency (responsiveness), server dependencies (resiliency), bandwidth and operation. The latter cannot be understated. There's a lot of really bad code running superfluous cycles on overclocked hardware, all bailed out by negative interest rates and cheap electricity. That era is over. As energy costs rise, so too does the cost to run bloat riddled with technical debt.

Pushing things to the interactive edge also delivers privacy and security gains over legacy designs. On-device data stores limit the fox to a single hen house, and decentralized learning models don't require you to expose your whole data stream in order to learn. Whatever you think about cloud, encryption or the clear net, there's no value in exposing data when you don't need to. Got a killer server-side geofencing app? How's it going to lock down a wayward device when the connection breaks? There's a practical side to this too. Governments are ramping up the regulatory pressure on data requirements, both at rest and in transit. It's the new oil after all. This is creating a regulatory minefield for global companies that send data across borders and even state lines.

It's clear from a bird's-eye where things are going. The information systems of the future will take more and more from the natural world, which is distributed and efficient in its energy expenditure. Tomorrow's systems will find no more value in needlessly pushing electrons around, anymore than nature finds value in pushing rocks uphill. Amazon in its wildest dreams couldn't scale twins in a data center the way nature can in a womb.

To be fair, legacy systems by way of their maturity offer a more robust feature set, but it's a trap. Given where we are in the evolution of the tech stack, investing in a legacy MDM (or any legacy tech solution for that matter) because of feature X, is like buying a steam locomotive in a maglev world, simply because you liked the stereo. That killer feature you bought might be the nail in your own coffin. As any architect worth their salt will tell you, modernizing a tech stack isn't as simple rewriting sections of code. You have to rip out the wooden tracks, redesign the cars, retrain the engineers, change the processes and culture--everything. A cursory Google of the term "Windows security flaws" will show you that hacks and patches will only get you so far, and you probably don't have Microsoft's resources.

The game is changing quickly and tech stacks will be split between the haves and have nots. Indeed we're already seeing it. And in the game of big data and AI, winner takes all. Can there be any other reason why Facebook still exists? Of course, don't take my word for it. I invite you to do your own research and figure it out for yourself. Because in times of serious inflection like these, vendor led is vendor dead.