As a CTO of a number of small startups, I am still struggling to understand what exactly AWS and other cloud providers give you to justify the markup.
And yes we’ve been heavy users of both AWS and Google Cloud for years, mainly because of the credits they initially provided, but also used VMs, dedicated servers and other services from Hetzner and OVH extensively.
In my experience, in terms of availability and security there’s not much difference in practice. There are tons of good tools nowadays to treat a physical server or a cluster of them as a cloud or a PaaS, it’s not really more work or responsibility, often it is actually simpler depending on the setup you choose. Most workloads do not require flexible compute capability and it’s also easy and fast to get it from these cheaper providers when you need to.
I feel like the industry has collectively accepted that Cloud prices are a cost of doing business and unquestionable, “nobody ever got fired for choosing IBM”. Thinking about costs from first principles is an important part of being an engineer.
And yes we’ve been heavy users of both AWS and Google Cloud for years, mainly because of the credits they initially provided, but also used VMs, dedicated servers and other services from Hetzner and OVH extensively.
In my experience, in terms of availability and security there’s not much difference in practice. There are tons of good tools nowadays to treat a physical server or a cluster of them as a cloud or a PaaS, it’s not really more work or responsibility, often it is actually simpler depending on the setup you choose. Most workloads do not require flexible compute capability and it’s also easy and fast to get it from these cheaper providers when you need to.
I feel like the industry has collectively accepted that Cloud prices are a cost of doing business and unquestionable, “nobody ever got fired for choosing IBM”. Thinking about costs from first principles is an important part of being an engineer.