Avoiding a Cloud Di-XaaS-ter

I still receive several emails a day on the wonders of cloud computing. The analysts and vendors fall over themselves to pronounce the Luddite nature of those companies who have not yet dumped their internal infrastructure or physical hosting for the ‘Shangri-La as a Service’ on offer. Many of these sales pitches now even cover the original concerns over security, privacy, security, compliance, security, IPR, ...did I mention security? 

To me it is not about being in front of or behind the curve. (Note that if you stay on the curve, you will probably go full circle and see the same solutions, rebadged, for all new problems. A number of times. As I have.) I’ve always been happy to use abstracted utility computing services and started my chequered career in IT running Payroll as a Service on an IBM mainframe for hundreds of companies. Fast forward to today, and we are all using the various cloud services on offer (Google, Amazon, Apple, Microsoft, Pirate Bay, - er, maybe not the last one anymore).

I also have no desire to continue to have my own server room taking up valuable real estate in the office, as I cannot get the economies of scale necessary to hit the proper/insane* (*Delete as appropriate) IT savings challenges. In this situation it is easy to listen to the ‘Snake-oil as a Service’ salesfolk, or more commonly be dragged in front of them by your nobbled CFO. So as well as the concerns mentioned above (did I mention... OK, sorry), I would recommend using the STIFF set of further criteria below to test the suitability of the service on offer, and the (in)credibility of the shiny suit Zoom-calling you.

• Standardised: One of the great myths of Cloud is that it is a set of standard commodity services that you can pick, choose, and swap like Lego bricks. I have yet to come across anything truly standardised in the cloud market; superficially a Wintel or Linux virtual server should be the same the internet over. So why do my team have to pore over the intimate details of server builds, SLAs, layered products, software versions, patch levels, etc., to get anywhere near a fair comparison? And don’t tell me PaaS is the answer as these are also subtly variant making pick ‘n’ mix tricky.
• Transformational: Ideally, you are not replacing like with like for your computing but will gain the many advantages touted for cloud – Mobility, Efficiency, Scalability, Speed. So why do many cloud migrations end up as a MESS?
• Insightful: Cloud seems a good place to create your data lake without worrying about space constraints, once you’ve figured out how to get your petabytes of historical data across the inadequate internet pipe put in by your network provider. Each IoT service can provide reporting or some sort of data feed, but it is proving nigh on impossible to get any sort of cloud-based distributed Master Data Management or enterprise-wide data model working effectively unless you stick to a single cloud provider and risk lock-in.
• Financial: If you are moving from the AMIGOS and expect to get a cheap deal ‘cause it’s cloud, then think again. You’re into a whole new opaque and complicated pricing methodology designed to suck you in and drain you dry.
• Flexible: Business flexibility is likely to be high on your shopping list from your users – if you are feeling a bit mean you can ask them to explain what they want, and then watch them squirm and bluster. Typically, they are actually looking for a quick reaction to events, which from my experience is rarely the rapid scaling that cloud does well. More often they want to launch a new product quickly or (de rigueur for 2020) cut costs aggressively. Unless you have been really smart (or lucky) in contracting for the XaaS, you will be disappointed by the difficulties in changing the service or driving cost down. One of the big gotchas of many SaaS and cloud contracts is the inability to scale down drastically mid-contract. And you thought upward-only rent rises was only a retail problem…

Don’t get me wrong, I see the move of IT to the cloud as both inevitable and desirable.  There are also some early mover benefits for some services that companies should explore, for example e-commerce & salesforce automation. But don’t expect it to solve your problems – if your strategy, process and data are screwed, putting your business up in the cloud will just highlight this to the whole world.

Remember, if you stick your head in the clouds, it just looks foggy.

John “Be Stiff” Moe