4 ms·

The figures presented here give a general idea but should be taken with care. A lot of these small virtual servers are packed into a single host. The actual pe

The figures presented here give a general idea but should be taken with care. A lot of these small virtual servers are packed into a single host. The actual performance you get depends a lot on how your neighbors are using the resources they share with you.

karmakaze 10y ago
I came to say exactly this. The pricing for resources chart is helpful. The performance results are based on too small a sample to be meaningful. It's really a measure of the load on the hardware by neighbouring accounts at the time of benchmarking.
equals2nine 10y ago
AFAIK, these $5 servers are VPS and the resources allocated are guaranteed so...how can neighbors affect the performance of your server? Are you saying they are actually shared hosting plans? If so, do you have any link/article on this?
- detaro 10y ago
  First: VPS vs. shared hosting doesn't mean (lack of) fixed resource allocation. Second, even if they allocate fixed resources to your VMs isolation is not that perfect, and you can see performance differences in things like memory access, storage speed, ...
- tyingq 10y ago
  Ignoring overselling for a minute, VPS isolation is not perfect. Some resources, like physical disk and network cards are almost always shared. Others, like CPU, memory bandwidth, etc, while theoretically partitioned, still suffer from some amount of contention.
- Lazare 10y ago
  > these $5 servers are VPS and the resources allocated are guaranteed so...how can neighbors affect the performance of your server Yes, they're VPSs, but no, a VPS does not provide the level of isolation that you're imagining. RAM is easy to offer guarantees about, CPU can be (but you can often burst over your guaranteed allotment, which means there's often high variability in performance, even if you always get what you're technically paying for), and disk IO is very hard (and you'll notice cheap VPSes don't even try and promise you any specific amount of IOPS). There's a million links discussing this, google "VPS noisy neighbor" or see, eg: https://www.liquidweb.com/blog/why-aws-is-bad-for-small-organizations-and-users/ https://www.liquidweb.com/blog/why-aws-is-bad-for-small-orga...

karmakaze 10y ago

I came to say exactly this. The pricing for resources chart is helpful. The performance results are based on too small a sample to be meaningful. It's really a measure of the load on the hardware by neighbouring accounts at the time of benchmarking.

equals2nine 10y ago

AFAIK, these $5 servers are VPS and the resources allocated are guaranteed so...how can neighbors affect the performance of your server? Are you saying they are actually shared hosting plans? If so, do you have any link/article on this?

detaro 10y ago
First: VPS vs. shared hosting doesn't mean (lack of) fixed resource allocation. Second, even if they allocate fixed resources to your VMs isolation is not that perfect, and you can see performance differences in things like memory access, storage speed, ...
tyingq 10y ago
Ignoring overselling for a minute, VPS isolation is not perfect. Some resources, like physical disk and network cards are almost always shared. Others, like CPU, memory bandwidth, etc, while theoretically partitioned, still suffer from some amount of contention.
Lazare 10y ago
> these $5 servers are VPS and the resources allocated are guaranteed so...how can neighbors affect the performance of your server Yes, they're VPSs, but no, a VPS does not provide the level of isolation that you're imagining. RAM is easy to offer guarantees about, CPU can be (but you can often burst over your guaranteed allotment, which means there's often high variability in performance, even if you always get what you're technically paying for), and disk IO is very hard (and you'll notice cheap VPSes don't even try and promise you any specific amount of IOPS). There's a million links discussing this, google "VPS noisy neighbor" or see, eg: https://www.liquidweb.com/blog/why-aws-is-bad-for-small-organizations-and-users/ https://www.liquidweb.com/blog/why-aws-is-bad-for-small-orga...