this post was submitted on 21 Aug 2023
2140 points (97.6% liked)
Technology
59695 readers
3538 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Has "cloud computing" ever been cheaper for most kinds of established businesses? Other than for some specific workflows, or very unpredictable workloads, the only cost-saving I've ever seen is avoiding the initial costs and avoiding the need for a real ops/obs team.
I can tell you at the enterprise level, Cloud services were absolutely pushed as a cost savings measure. All the math in the world can't save you from a determined C-suite, however.
We just finished our migration to the Cloud after 3 long years of effort, and while we are saving about ~2MM/mo in data center costs, our opex spend is up by around 2.5MM/mo YoY, not including all the Cloud-centric new hires.
Are you saying they already (over)spent in unrelated opex areas the savings from going to the cloud? I'm unclear if you're saying it's a consequence of the move to the cloud.
Our run rate is roughly 2.5MM more per month than what we were spending to operate two whole-ass data centers.
Hope that clarifies it a bit.
Initial and operational costs are huge if you are a small company of ~20 people. At least in this case the promise of cloud is achieved - bringing the economies of scale down to individuals and small companies.
Sure if you have 10k employees it makes no sense, you have enough resources for these same economies of scale to be possible inside your company.
You don't even need 10k employees, I see it make sense with ~450 employees if you also have a decent IT team and funding. The issue is most companies can't see the need to keep things they own up to date - there's always a temptation to "just put it off a year" to make the budgets look better, till they hit near catastrophe with being 5+ years beyond reasonable. The cloud "forces" them to put in update, maintenance, employee overhead etc up front and forever. They just pay a premium for that service IMO.
I used to think it was kind of stupid, but then I realized - companies hire consultants at exorbitant rates to help them do things they don't have the in house skills for - so really - building that into the overall cost might still be a wash. The expensive part of Cloud IMO turns out to be needing training, consultants or new employees with different skills to manage it, which all charge more than traditional on prem because cloud is still the current ?fad?. And the unseen costs of screw ups by the cloud provider themselves losing data, being down, or having a security breach that affects you - and you're completely out of the picture with remediation or even knowing what might be a risk.
And easy scaling options.
As someone who does DevOps for a living, The scaling options are really what make cloud semi-affordable and useful for enterprises. Not to mention the “I don’t have to waste engineers doing menial upkeep” (aka, managed services means there’s a good amount of “not my problem”). The other part that’s a huge savings is being able to use things like terraform or pulumi to quickly deploy, destroy and redeploy for test environments and dr.
I can completely redeploy an enterprise scale website in hours, code and data deployments included.
Things are crazy busy because you ran a sale or ad during the superbowl, scale it all up in seconds. Want to test something or run a dev environment that people don’t use regularly? Only spin it up when they need it.
We power down all dev and test environments every night and weekend. Some only spin up on demand. Saves tons on capital and nominal run rate.
Yep, that's what I meant by "unpredictable workloads".