Getting our applications out of the cloud provided the main celebration for our exit, but seeing the actual spend tumble is the prize. See, the only way to get pricing in the cloud down from obscene to merely offensive is through reserved instances. This is where you sign up for a year or more in advance on a certain level of spend. Th...
Having your compute in “the cloud” doesn’t remove the need for a good backup strategy, it just changes how it works. Yes, disaster recover for natural disasters should be easier (OHV’s fire showed that this may not always be true). But, that doesn’t cover cases like ransomware, insider threats, data mistakes or any other case where data is corrupted/modified by mistake. You still need a plan for these cases. And cloud based backups actually make a lot of sense.
But, just because you put your backups in the cloud, doesn’t mean that your compute should be there as well. There is an advantage that your Time to Recovery is likely lower with both backups and compute in the same cloud. But, is that worth the ongoing cost of running your compute in the cloud? That needs to be considered separately. You also need to consider the cost of running on-prem versus in the cloud. If you have fairly predictable, static loads, it may be cheaper to buy and run servers yourself. For hard to predict, elastic loads, cloud may make more financial sense.
As others have said before, there was a period where companies were just going to the cloud for the sole reason that it was the popular thing to do. For some it actually made financial sense. For some, it didn’t. The OP’s article seems to be the latter.
Exactly. Use cloud for off-site backup and things that need flexibility.
You don’t need any of that to run a basic website. You can almost use an old laptop or PC for most static applications.
The cloud isn’t just for storage or compute. There are a number of managed services that let you build a full application by snapping together lego building blocks.
For example, pop together a REST API handler, an auth service, a few functions-as-a-service, a database, and a storage service. Then add a static website server. Throw a CDN in front. You got yourself a dynamic application service that can be accessed globally for a few pennies and can scale up and down without you doing anything. Add multi-zone support and auto-DNS failover and you’ve got a production quality scalable, resilient back-end, for both web and mobile. When it’s not being used, it costs very little and when it goes big, hopefully it means you’re doing well. Wrap it all in an infrastructures-as-code script and you can bring all this up in 30m.
To host all that in-house, you would have to buy a lot of equipment, stage it, manage it, add cooling, electricity, security patches, upgrades, security, etc. Now you have part of your business just doing all this instead of focusing on what you do best. I won’t bother going into the tax implications of capex vs opex.
This, is what the cloud sales people call ‘undifferentiated heavy lifting.’ There are reasons to have on-prem hardware. For a lot of applications though, it makes more sense to let someone else take care of all that infrastructure cruft.