Last weekend I had a chance to talk with a person who (finally) gave me confirmation that I am not alone in my opinion on the topic of "sending everything to the cloud".
The subject is more complex than it may seem, but I would like to ask those who present cloud technology as the cure for all evils:
How aware are you of going to the cloud? And how bound will you be to your service provider once it's done?
Let's clarify.
"Let's go on AWS, because there's this and this and that, which are REALLY COOL and allow us to do things that would otherwise cose too much".
That's true. But what happens when the provider you chose decides to completely abandon a solution? Which could also just mean they will no longer update it.
How bound to the provider is the solution you're planning to implement?
And what if you just want to change your provider?
Having something on cloud doesn't represent a problem, but having everything in full outsourcing may not be a problem in the middle term. However, the logic should be - in my opinion - to adopt "carrier neutral" solutions (it would be more correct to say "provider neutral", but I have a personal attachment to the terms), those solutions which don't bind you to a single CSP (Cloud Service Provider) but rather allow for the possibility to change provider to a closer or, why not, cheaper one, if necessary.
So, if "let's put everything on Amazon S3" becomes an economic abyss because the service costs as much as it's worth (and it's worth a lot), having an exit strategy may become a necessity.
And we haven't even touched the data topic yet.
There's only one answer.
Knowing the technologies and the market.
I don't mean you need to know how to write a MQ (Message Queuing) system from scratch, but knowing how to implement (or adapt) your code to move to a different carrier is necessary.
The Hosting services market has changed radically in the past few years. There are providers offering exceptional services for extremely fair prices, and service providers offering pretty much everything, but could cost a big part of the budget of a start-up company.
Maybe, and this is also my personal opinion, trusting a Cloud Service Provider which tends to take everything in while limiting the possibility of movement, could become a problem within 18-24 months, as you would emd up bound to the single provider, with little knowledge of their architecture.
Mind Your Own Business
Then there's the data aspect. I'll admit the sentence is stuck in my head, but that's because I find it a perfect fit for several topics that people are talking about a lot these days, but haven't really changed.
The topics could be privacy and GDPR, but I trust that I can take it from a higher level:
your company data is yours and only yours.
Documents, proposals, projects, invoices, contracts. Which parts of this data can be lightheartedly handed over to a cloud service? Which parts cannot?
Data protection and information security in the "cloud service providers" era could become a serious issue within a few years. But if my calculations are incorrect (although recent events seem to agree with me), it may even only take a few months.