It’s that time again!
This month I decided to compare some higher end 16GB servers instead of the
usual $5 to $10 dollar instances. Each costing only $80 per month, these servers
have a ton of bang for the buck and should be more than enough for even a high
trafficked site.
I took a bit of time today to work through the issues I was experiencing with
the MySQL benchmarks, and I am happy to report that they are included once
again!
As always, the following stats are an average across 3 server instances from
each provider.
Overview
DigitalOcean | Linode | Vultr | |
---|---|---|---|
Location | New York 1 | Newark, NJ | New York (NJ) |
Operating System | Ubuntu 18.04 LTS | Ubuntu 18.04 LTS | Ubuntu 18.04 LTS |
RAM | 16 GB | 16 GB | 16 GB |
CPU | 6 Cores | 6 Cores | 6 Cores |
SSD | 320 GB | 320 GB | 200 GB |
Transfer | 6 TB | 8 TB | 5 TB |
Note: Thanks to everybody that reached out over the transfer allotments
being swapped between Linode and DigitalOcean. Things are sorted now, so again,
many thanks 🙂
CPU Info
DigitalOcean | Linode | Vultr | |
---|---|---|---|
CPU MHz | 2294.61 | 2533.33 | 2593.90 |
Cache Size (KB) | 25344.00 | 16384.00 | 16384.00 |
BogoMips | 3059.67 | 3533.00 | 3460.00 |
CPU
DigitalOcean | Linode | Vultr | |
---|---|---|---|
Events per Second | 1087.14 | 794.31 | 1144.08 |
Minimum (ms) | 0.85 | 1.21 | 0.85 |
Average (ms) | 0.92 | 1.26 | 0.87 |
Maximum (ms) | 3.16 | 2.17 | 2.07 |
Memory (Read)
DigitalOcean | Linode | Vultr | |
---|---|---|---|
Ops per Second | 4367899.44 | 2428632.70 | 4522785.19 |
Minimum (ms) | 0.00 | 0.00 | 0.00 |
Average (ms) | 0.00 | 0.00 | 0.00 |
Maximum (ms) | 0.19 | 0.22 | 1.19 |
Memory (Write)
DigitalOcean | Linode | Vultr | |
---|---|---|---|
Ops per Second | 4357797.53 | 2330957.50 | 4529815.35 |
Minimum (ms) | 0.00 | 0.00 | 0.00 |
Average (ms) | 0.00 | 0.00 | 0.00 |
Maximum (ms) | 0.06 | 1.82 | 1.77 |
File I/O
DigitalOcean | Linode | Vultr | |
---|---|---|---|
Reads per Second | 1921.01 | 4384.63 | 6996.74 |
Writes per Second | 1280.67 | 2923.09 | 4664.50 |
Fsyncs per Second | 4092.32 | 9348.55 | 14920.48 |
Minimum (ms) | 0.00 | 0.00 | 0.00 |
Average (ms) | 0.14 | 0.06 | 0.04 |
Maximum (ms) | 24.59 | 26.70 | 5.54 |
MySQL
DigitalOcean | Linode | Vultr | |
---|---|---|---|
Transactions per Second | 2397.33 | 3210.67 | 6024.00 |
Queries per Second | 47946.67 | 64213.33 | 120480.00 |
Minimum (ms) | 1.81 | 2.07 | 1.19 |
Average (ms) | 4.27 | 3.18 | 1.66 |
Maximum (ms) | 39.83 | 18.06 | 36.35 |
Speed Test
DigitalOcean | Linode | Vultr | |
---|---|---|---|
Distance (km) | 2433.66 | 2373.19 | 1871.87 |
Latency (ms) | 47.119 | 65.218 | 45.610 |
Download (Mbit/s) | 1329.49 | 376.57 | 1446.49 |
Upload (Mbit/s) | 468.86 | 325.12 | 514.80 |
Apache Benchmark (against nginx
on the servers)
DigitalOcean | Linode | Vultr | |
---|---|---|---|
Requests per Second | 220.20 | 194.34 | 200.11 |
Time per Request (ms) (mean) | 2270.79 | 2594.23 | 2505.67 |
Transfer Rate (Kbytes/sec) | 184.72 | 163.03 | 167.87 |
Conclusion
While each server had comparable price, RAM and CPU cores, Vultr was far behind
DigitalOcean and Linode in regard to disk space offered. Both DigitalOcean and
Vultr trailed behind Linode’s transfer offering.
Even with Linode and Vultr having comparable CPU speeds, Linode fell short in
comparison when it came to the number of CPU events processed. Linode fell even
shorter in both memory reads and write as well.
Vultr really shined for file I/O with DigitalOcean coming in more than 50% the
speed for reads and writes. Vultr continued to shine on MySQL with both Linode
and DigitalOcean coming in at least 50% the speed.
Speaking of speed, Vultr and DigitalOcean’s network speeds were comparable and
blew Linode out of the water. With that, nobody could touch DigitalOcean serving
up requests from nginx.
As always YMMV and you should evaluate your own personal needs when picking a
hosting provider. Vultr shined quite a bit on these benchmarks but the limited
SSD and transfer allotment has me leaning towards DigitalOcean as this month’s
pick.
Also want to mention, I played with Linode’s beta control panel during this
month’s benchmarks and was quite impressed. It shows me that even though Linode
may be falling behind a bit, there’s a good chance they still have some tricks
up their sleeve for the future.
If you found this post helpful and would like to help support me while signing
up for your shiny new server, please consider using my referral links below.
Now that I got some things sorted with the benchmarks, I’m hoping to include a
new provider next month… maybe 😉