Skip to content
Register

YABS on 50 different $10/year plans, the variance is absurd

VPS Hosting by petra 29 replies 4.8K views
petra

Member OP

#1

We ran YABS on 50 different $4-60/month plans across

  • Contabo
  • OVHcloud
  • Vultr
  • Hetzner
  • InterServer
  • Hostinger
  • KnownHost
  • RackNerd
  • GreenCloudVPS
  • HostHatch
  • CloudCone
  • Leaseweb
  • Time4VPS

The variance is absurd. Some NVMe plans score worse than decade-old SSD. I can share the dataset if anyone wants to dig. Personally I expected tighter clustering, but generational hardware splits are messy.

lucgone

Member

Le Baguette

#2

Poggers for doing the lords work, servr lottery is real, that one Leaseweb s3rver literally had hdd from 2014, vibe check failed, based that you exposed this

Vive la résistance... électrique
minh1987

Member

#3

petra said:
I can share the dataset

Can you help me,very interesting result,the ser-va with very good disk is very fast,okay okay okay,can you help me how to check my own ser-va,very thank you

phở at 3AM, deploy at 4
RAJ

Member

#4

Guys any leads on which provider is most consistent? Looking to add white label options, margins on these cheap plans are thin already. Recieved some offers from their resellers but hard to tell 🙏

your margin is my opportunity
marcus_qc

Member

#5

petra said:
Some NVMe plans score worse than decade-old SSD.

Hey folks, super stoked someone finally benchmarked this properly. We had a bad batch at InterServer last year, gonna own that, some E5 v2 nodes slipped through. Fixed now. Your data shows it, we were the overprovision outlier in Q3. Replaced hardware already, happy to comp tests.

42U and still growing
DA

Member

#6

The server lottery is real lah, had a 2014 hdd surprise too, never again

minh1987

Member

#7

daniel34 said:
Can you help me how to check my own ser-va

Very thank you for wait, I find it, YABS is on github, I run it now on my Vultr Singapore, result is okay but not same as petra test, maybe different node

phở at 3AM, deploy at 4
torontoeng

Member

infrastructure as code

#8

petra said:
I can share the dataset

Please do. I want to see the Hetzner Falkenstein vs Helsinki split. My home lab ping to Helsinki is 110ms, Falkenstein 125ms, but I care about disk consistency more.

petra

Member OP

#9

torontoeng said:
Please do.

Uploading now. 47MB CSV. The Hetzner split is there, two generations in Helsinki, one older. Falkenstein is actually tighter than I expected.

RAJ

Member

#10

marcus_qc said:
Replaced hardware already, happy to comp tests.

Good to know. What about Hostinger? Their Mumbai node was all over the place in my own small test. Some NVMe, some not, same plan name. Hard to sell that white label.

your margin is my opportunity

Post a reply

You need an account to reply. Log in or register to join the conversation.

Post reply Preview Save draft