Although AWS and similar solutions offer a low cost of entry, with typical systems involving 100s of machines, maybe the cost of the cloud is not as trivial as we first thought…
I was initially prompted to write about this when I noticed developer laptops specifications being drastically upgraded to support 10+ virtual machines in order to support running a small part of a microservices architecture under development. Yes, a developer could run one part easily on a normal laptop, but in order to test out even simple scenarios, 10 different VMs needed to be configured and started up. In contrast a Elixir/Phoenix/PostgreSQL development environment can run on a relatively cheap, low memory laptop.
Marcus Ranum recently wrote about being Lost in the Clouds
“In its most basic form, cloud computing allows you to transfer some risks around, and that’s it. It saves you money if you choose well and can aggregate services with other customers and avoid lock-in, and it allows you to fire those pesky system administrators who used to manage your storage array. Instead of capital expenses for salary and desks and hard drives, you can pay more for something you don’t own which is slower and out of your direct control.” [Emphasis in the original]
Andreessen Horowitz in The New Business of AI, –hat tip to Scott Locklin – have reported that
“… these forces contribute to the 25% or more of revenue that AI companies often spend on cloud resources. “
The problem is that running an AI model typically takes a very large data set and a lot of processing, both of which are expensive in the cloud. While these expenses are operating costs, it can seem to be cost effective, but as Ranum points out, it might be that purchasing machines and hosting them onsite is a cheaper and faster solution.
My take is that the cloud is reinventing the big-iron era of the mainframe, where companies end up paying by the hour for processing time, storage and network bandwidth. Yes, there may be some circumstances where this makes sense, but when you need to start hiring administrators to configure and manage your cloud, maybe it is time to to consider the alternative.