Comparison: Vultr vs. Digital Ocean vs. Linode vs. SSD Nodes

Comparison: Vultr vs. Digital Ocean vs. Linode vs. SSD Nodes

Posted by Joel Hans on June 28, 2017

Whether you’re looking for your very first VPS or you have so many you’ve lost count, it’s understandable that you’d want to do your research before shelling out for a new server. It’s a big investment, after all, particularly if you want to snag the best deals by paying annual instead of monthly (or even hourly).

Sadly, SSD Nodes often gets left out of these round-ups, mostly because it’s much easier to focus on just the biggest players while leaving the up-and-coming competition out of the picture. Because we keep seeing this comparisons come and go without a single mention of our name, we though it was about time to put our own together.

Plus, we’re not afraid to punch-up against our biggest competitors.

Here’s a couple of things to keep in mind moving forward:

  • All of these prices are for unmanaged servers—that means you install and maintain everything—and are for shared resources (aka not dedicated resources).
  • All the providers offer full root access.
  • There’s variability on which options are available, but all the providers let you choose from a variety of Linux distributions, with Ubuntu, Debian, and Cent OS being the most popular.
  • This comparison doesn’t focus on those extra-small servers (we’re talking 512MB of RAM kind of small), but rather the types of servers that people use to get actual work done. In this case, they all come with 8GB of RAM.

Instead of moving through each provider and trying to catalog the more subjective measures—who has the best technical support, who has the prettiest dashboards, which has certain extra features—we though it best to hone in on the facts: what each provider offers in terms of RAM/disk/transfer, which servers benchmark the highest, and which gives the most performance for your buck.

The basics: what each plan offers

Let’s get a baseline for what we’re looking at with the 8GB servers from these various providers.

8GB plans Vultr Digital Ocean Linode SSD Nodes
RAM 8GB 8GB 8GB 8GB
CPU 4-core 4-core 4-core 4-core
Disk 100GB SSD 80GB SSD 96GB SSD 40GB SSD
Transfer 4TB 5TB 4TB 4TB
Price $40/mo $80/mo $40/mo $9.99/mo

Time for some benchmarks!

Now, we took a slightly different approach to benchmarking than you might see in other comparisons. Instead of buying a server at each of these providers and throwing some basic benchmarks at it, we instead relied on scores uploaded to ServerScope. Why? We felt as though these scores, submitted by actual users at each provider via a script that performs exactly the same benchmarks, could provide a more objective result than if we did the benchmarks ourselves.

ServerScope benchmarks look at a few important aspects of server’s performance:

  • A speedtest benchmark to see how quickly files can be uploaded to the server.
  • A bandwidth test to measure download speeds from a major CDN.
  • Disk speed using dd.
  • Disk I/O with fio.
  • And a comprehensive Unixbench score, which aggregates the speed from a number of different areas into score that can be easily compared between servers.

Again, using the 8GB plans as our baseline (and the 3GB at ChicagoVPS), we found ServerScope benchmarks for each of the plans in question.

  Vultr Digital Ocean Linode SSD Nodes
Disk read 6144 MB/s 2839 MB/s 5538 MB/s 13236 MB/s
Disk write 1444 MB/s 679 MB/s 1307 MB/s 2999 MB/s
dd (64K/32K)/(1M/2048) 388 MB/s / 389 MB/s 188 MB/s / 225 MB/s 683 MB/s / 713 MB/s 1.5 GB/s / 1.5 GB/s
Read IOPS 1610649 726797 1451827 3469721
Bandwidth 799.84 Mbit/s 848.07 Mbit/s 654.99 Mbit/s 28.93 Mbit/s
Speedtest 529.32 Mbit/s 396.69 Mbit/s 370.61 Mbit/s 35.95 Mbit/s
Unixbench 3957.9 2298.1 3555.9 5186.0*

Now, we noticed that our network scores weren’t great via the one user-submitted benchmark. This being our comparison, we decided to run an additional test to see what we could come up with. And, one test later on the same spec’d server as the other benchmark, we ended up with some more reasonable (and dare we say realistic) results: bandwidth at 1305.48 Mbit/s, and speedtest at 107.95 Mbit/s.

And here’s are the links to each test result:

Calculating the bang for your buck

This is where things get interesting, if you ask us. If you’re after raw performance, of course you can pay hundreds per month for a top-of-the-line dedicated server with all the speed you’ll ever need—at least until you outgrow it, of course.

But what if you need to save some cash now, and still get a solid baseline performance while you grow? That’s where not only price, but also value, comes into play.

RAM/dollar

All of the plans we're talking about here come with 8GB of RAM, but that doesn’t mean that each GB of RAM in each of them costs you the same. If you're looking for your money to go its absolute furthest, this is an important figure. The chart below outlines how much RAM (in gigabytes/GB) you for a single dollar with each of the providers.

 

Performance/dollar

Just as with RAM/dollar, we think it's important to examine how far your dollar goes when it comes to raw computing performance. That might be a major factor when deciding which provider to go with, if you want to move quickly but don't want to pay an exorbitant amount per month to make it happen.

The chart below outlines how much Unixbench performance you get for every dollar you spend.

Wrapping up

Obviously, there are a lot more factors that go into your VPS choice than raw speed and price—some want the absolute cheapest offering available, and others want more unique features, like load balancers. We wholeheartedly believe that every potential VPS user should take time for market research before jumping into any VPS purchase. The last thing you want is to be surprised by

Other points to consider:

  • What platform does the provider use? Xen, Virtuozzo, KVM?
  • Which platform features do you actually need? Do you need to run that obscure Linux distribution, or is Ubuntu 16.04 fine?
  • Do you need partially- or fully-managed service?
  • What do others think about the service?
  • Has the provider been around for a while?

We hope this helps address at least some of the issues around picking a VPS provider. If you have any questions for us, feel free to ask in the comments!

Topics: benchmarking

New Call-to-action

Subscribe to our email updates