Cloud VM Resource Allocation

The cluster discusses how cloud providers like DigitalOcean, Linode, and others allocate CPU and RAM resources to virtual machines, focusing on shared vCPUs, dedicated RAM, overcommitment, and performance implications compared to dedicated instances.

📉 Falling 0.5x DevOps & Infrastructure
3,212
Comments
20
Years Active
5
Top Authors
#1593
Topic ID

Activity Over Time

2007
7
2008
13
2009
82
2010
96
2011
98
2012
148
2013
191
2014
147
2015
151
2016
150
2017
216
2018
177
2019
231
2020
227
2021
277
2022
256
2023
271
2024
233
2025
227
2026
18

Keywords

RAM VCPU CPU TBH AMI K8 AWS EC2 linode.com VPS vm cpu ram vms virtual machine cores shared storage disks

Sample Comments

zagi Apr 29, 2013 View on HN

RAM is dedicated on DigitalOcean, CPU's are not - the virtual cores are shared to some extent on almost any cloud provider if you do the math.

andersmurphy Oct 17, 2025 View on HN

Yeah shared vCPU can be really bad.

ChuckNorris89 Jul 15, 2020 View on HN

Lots of VMs. VMs don't need much CPU time but oh boy do they need the RAM.

joeyrobert Apr 19, 2016 View on HN

Do it, they're incentivizing you to bring up 8 VMs or 1 really big VM.

copperx Apr 18, 2017 View on HN

What if you were running a few virtual machines to develop (very common workload)?

irons Apr 15, 2011 View on HN

Doesn't seem likely. VMs impose limits on resource consumption (though Rackspace allows CPU bursting, from what I understand). It's not like the old days where you could get lucky on a sparsely-populated shared server.

threeseed Jan 29, 2014 View on HN

Tell me you're joking ?You're actually comparing virtual CPUs between two providers.

manigandham Dec 13, 2016 View on HN

Citation needed for what? its common knowledge that the major clouds VMs with lower performance and higher cost in exchange for flexibility.

applebug60 Feb 6, 2013 View on HN

They're both using the same virtualization. What's your use case that requires that much CPU? I'd bet spot instances would be a better fit in most cases.

NBJack Feb 11, 2025 View on HN

Not to my knowledge, no. If it's a virtualized server, maybe this works, but to my knowledge most companies try to maximize their hardware usage through a variety of slices/sizes based on underlying resources. I am not a datacenter expert however.