Hacker News Reader

Comment by paulddraper | Hacker News Reader

This is absolutely correct (and gets more correct every year).

A m5.large (2 vCPU, 8GB RAM) is $0.096/hr. m5.24xlarge (96 vCPU, 384GB RAM) is $4.608/hr.

Exactly 1:48 scale up, in capacity and cost.

The largest AWS instance is x2iedn.32xlarge (128 vCPU, 4096GB RAM) for $26.676/hr. Compared to m5.large, a 64x increase in compute and 512x increase in memory for 277x the cost.

Long story short.....you can scale up linearly for a long time in the cloud.

View on HN

Except, it’s hard to do vertical scaling dynamically.

That kind of kills this whole argument except in certain pretty unusual cases.

I think for most use cases overprovisioning a single instance to meet max denand will still be cheaper than dynamic provisioning with the synchronization overhead of a distributed system.

> Except, it’s hard to do vertical scaling dynamically.

Easiest to just create a new instance scaled up vertically and swap it over.

How many times do you need to scale? In the usual case.

What’s the “usual case”? We scale up and down thousands of times a day.

Yeahz but the argument being made is that there's overhead in that, primarily in the extra Dev time needed to build such a system, and the extra failure points introduced requiring more maintenance effort.

Those two items can be substantial (5x headcount), or they can be merely "standard" (2x headcount).

They're never zero extra headcount.

> What’s the “usual case”?

Whatever isn't the "unusual cases" you mentioned.

> We scale up and down thousands of times a day.

Sounds unnecessary for most cases.

> A m5.large (2 vCPU, 8GB RAM) is $0.096/hr. m5.24xlarge (96 vCPU, 384GB RAM) is $4.608/hr. Exactly 1:48 scale up, in capacity and cost.

Only if you're paying on-demand prices. If you're paying spot prices, the spot market is much less linearly correlated.

Reply on news.ycombinator.com