I saved $183M dollars by not moving to the cloud(pauldambra.dev)
pauldambra.dev
I saved $183M dollars by not moving to the cloud
https://pauldambra.dev/2023/03/the-cloud.html
7 comments
It seems to me like you are considering the benefits of moving to cloud (which are real!) but not the costs? If it's part of your personal job responsibilities to maintain physical infrastructure, and if by moving to the cloud you get someone else to take on that aspect of it, but otherwise keep your job, then aren't you always going to prefer cloud?
> considering the benefits of moving to cloud (which are real!) but not the costs
I was being very tongue in cheek :)
> if by moving to the cloud you get someone else to take on that aspect of it, but otherwise keep your job, then aren't you always going to prefer cloud?
Quite possibly, yes. Paraphrasing Dan North: "Infrastructure is like surgery. If you need surgery, you really want to have surgery. But you always want to have the smallest amount of surgery possible."
Having a rack in a basement or co-lo might be "cheaper" for you. It might end up being a pain. The BMC infrastructure was relatively simple. We were early adopters of EC2 (so the financials _might_ have changed since then), we saved about 60% in explicit cost by shifting from co-lo to EC2. And my time was freed up because I didn't have to plan for or handle anywhere near as much infrastructure work.
We didn't make money directly from infrastructure so it always made sense to reduce the amount of infra I had to do because it increased the time I could spend closer to revenue generation
I was being very tongue in cheek :)
> if by moving to the cloud you get someone else to take on that aspect of it, but otherwise keep your job, then aren't you always going to prefer cloud?
Quite possibly, yes. Paraphrasing Dan North: "Infrastructure is like surgery. If you need surgery, you really want to have surgery. But you always want to have the smallest amount of surgery possible."
Having a rack in a basement or co-lo might be "cheaper" for you. It might end up being a pain. The BMC infrastructure was relatively simple. We were early adopters of EC2 (so the financials _might_ have changed since then), we saved about 60% in explicit cost by shifting from co-lo to EC2. And my time was freed up because I didn't have to plan for or handle anywhere near as much infrastructure work.
We didn't make money directly from infrastructure so it always made sense to reduce the amount of infra I had to do because it increased the time I could spend closer to revenue generation
In a small organisation without enough staff to cover revenue generating roles I'd expect the cloud almost always makes sense. The opportunity cost of giving up revenue generation to save on infra costs is unlikely to work out. Sadly I suspect that in a large organization the people who look after hardware get laid off rather than moved closer to revenue generation!
Since folk like DHH and Ahrefs are posting their savings from running workloads outside of the cloud, I thought I'd join in
Where did DHH post his?
This post is just someone felt personally attacked for no reason and wrote an emotional text with no context.