Cloud may be the future, but it ain't all sunshine and rainbows Yes, cloud might be the future but what truths lie hidden beneath this rock of certainty? You've heard the hype, pros and cons, but there's plenty the average cloud user may not have considered in the clamour to get up there. Our company recently heeded the cloud's call, and this is what we discovered. Where cloud excels is …
When was cloud ever about cost savings?
Outsourcing of any sort is about turning capex into opex i.e. spreading the cost over time.
With cloud in particular it's also about scalability. If you are planning on growing fast you can start small at very low cost, with the costs only scaling as your business scales. When you're building on-prem infrastructure you have to predict the future to a much larger extent. If you think you're going to grow from a 1000 to 1000000 customers in 8 months, you need to build much of the infrastructure for that up front. And what if your predictions are wrong?
Doing cloud properly is never going to be about saving money.
This I agree 100% I jumped from enterprise IT engineer to a small startup developer/dba/cloud admin and the difference is really clear.
For my old employer cloud didn't make financial sense or wasn't technically viable idea outside of the web presence (web/mobile apps) and some things like CRM, email, and VOIP provided as SaaS solutions.
Environment consisted mostly old enterprise stack where even if the software would scale as services licensing definitely wouldn't. Plus existing DC + staff made running own apps rather cheap, especially where old servers could be refurbished to new use.
Looking at working at a startup first there:
1. There is no DC & staff to run it cost even setting up colocation side would be substantial.
2. You likely have no idea of the real scale of the workload you end up running even within a 3-month timeframe.
3. There is no legacy. This means you can really think everything as a service engineering systems that really run well in the cloud taking the full advance scaling per demand, PaaS services etc.
4. Building and running bought software & hardware wouldn't be financially sane. Startup/new company is always heavily CapEx limited. Most don't care what something costs in 3 years timeframe as there is no knowledge if the company is even going to exist in 3 years. Most important is to fix cash flow now. Meaning not tying capital upfront on anything if not absolutely required.
This said even for old enterprise cloud will make sense for innovation projects like industrial company expanding to IoT space or traditional marketing company expanding to data analytics and digital content delivery.
It's just that cloud journey should never start by migrating ERP to cloud in order to save cost. First, because cost savings are unlikely to materialize secondly project has a pretty good chance of failing.
And the final thing the fact that cloud providers limit your configuration, protocols and local storage is actually a good thing. Fact that you simply cannot use some nasty hacks used in enterprise space and call them "integrations" actually leads to more reliable more maintainable systems.
Keep these articles coming, Reg editors. I am showing them to as many people as I can.
It feels as if peak cloud hype is passed. We're beyond the unicorn stage and now there are enough people having rolled up their sleeves and put their hands in the gunk to come back and tell us their cautionary tale.
I'd gladly pay 'the proce' if I'd knew the price.
Now, moving to something else the article says something that is not often realised is that every byte of data that leaves the cloud provider's data centre will be billed' but stops right there. Where's the 'will be inspected, mined for any info, even sold to interested parties'.
Cloud is nice, as long as is under your control. As soon as the data leaves your control the game is over. Your data doesn't belog to you anymore, it belongs to the cloud provider and can (and most surely will) be hold hostage every time the price will increase.
Cloud is nice.
It allows you to not deal with most of the plumbing
In my experience, you still have to deal with the plumbing, except now you have bits which have metric connectors AND bits which have imperial connectors, and you have to hook it up to something that Heath Robinson* rejected for using pre-industrial age fittings.
* Rube Goldberg for those of a left-pondian
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
