AWS Lambda was already serverless, now it can be x86-less too Amazon Web Services has made it possible to use its home-baked, Arm-powered Graviton2 CPUs with its Lambda serverless functions. The server-renter on Wednesday slipped out news that Gravitons are an option alongside its x86 offerings from Intel and AMD. It will come as no surprise that AWS reckons its own silicon slays on …
Of course in reality 'serverless' is not serverless. There must be servers for the transactions to take place. It's just that the servers are transient. Obviously that can be an advantage to the provider as it reduces the amount of idle processing power across the large infrastructure (most transactions being bursty). I'm not sure what specific benefit that offers the user though.
Serverless benefits the user over IaaS because it actually delivers the cloud premise of freeing them from managing OS instances. IME, that's where most of the maintenance effort goes (more than the hardware/firmware/hypervisor aspect of servers that IaaS covers, in fact).
It's generally more effort to adopt, and less interchangeable with other deployment types, but it certainly makes a difference to the amount of non-application operations effort required. It's probably what I'd choose if starting from scratch, in fact.
I wonder what kind of use cases it fits?
I mean, most of the time it is still cheaper to provision beefy dedicated servers and hire a DevOps team full time. Plus you have a total control over your infrastructure and you don't depend on refreshing the status page of cloud provider when things go south.
Is it though?
You leave a server running with AWS over a month with a set of Devops monitoring it can you realistically say that that is going to be cheaper than using a serverless approach?
If you have a Lambda that acquiesces between invocations it will always be cheaper.
To prove the point I wrote a dumb IG trading bot based solely off the free tier Lamda - yep - I made money from nothing...
I will accept that you may need to spend some time/money to refactor parts of your code base - but thats a one off cost.
Generic solutions degrade as your requirements become more specific. AWS Serverless becomes exponentially complex when it comes time to scale-up the features. Even AWS hosting has a point where all of the complex costs and bewildering performance bottlenecks are more difficult to work within than running your own hardware.
If you have a function which runs only periodically, you don't need an instance running while it's inactive, potentially saving money. Because it's serverless (from the consumer point of view), you don't have the worries of patching, OS maintenance, security etc and it's highly available because AWS will just run it on whatever bit of tin is available at the time you need to run the function.
It's a niche usage, but there are some white papers/blog posts where you can effectively run an entire website using Lambda functions, so there's some useful capabilities in there.
This reminds my of the old Intergraph CAD computers that ran on their own custom designed Clipper chips. Well, at least until it switched to microsoft os. You know, the company that went from one of the top 100 software companies in the world in 2008 to being sold to Hexagon by three quick buck private venture firms.
The Register - Independent news and views for the tech community. Part of Situation Publishing
Biting the hand that feeds IT © 1998–2022
