Changing it up a bit this month. No new providers have been added, but by
request, I’ve moved the entire operation overseas to test out data centers in
London.
Incidentally, every provider covered has a data center that’s actually in
London so there’s no concession on location by having to pick a data center
that’s near a certain location, as I have to when comparing US-based data
centers.
As per usual, I’ve spun up 3 instances with each provider, all running Ubuntu
18.04 LTS, and have averaged the results together.
Another request I didn’t quite get to, but am hoping to include next month, are
benchmarks of NoSQL databases / caching stores, specifically MongoDB and Redis
(and maybe even Memcached).
Overview
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Location |
London 1 |
London, Zone A |
London, UK |
London #1 |
London |
| RAM |
1 GB |
1 GB |
1 GB |
1 GB |
1 GB |
| CPU |
1 Core |
1 Core |
1 Core |
1 Core |
1 Core |
| SSD |
25 GB |
40 GB |
25 GB |
25 GB |
25 GB |
| Transfer |
1 TB |
2 TB |
1 TB |
1 TB |
1 TB |
CPU Info
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| CPU MHz |
2329.74 |
2400.14 |
1999.99 |
2999.98 |
2394.45 |
| Cache Size (KB) |
22016.00 |
30720.00 |
512.00 |
16384.00 |
16384.00 |
| BogoMips |
3129.67 |
3200.33 |
2666.67 |
3999.67 |
3192.33 |
CPU
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Events per Second |
822.67 |
862.97 |
1257.41 |
1064.70 |
792.91 |
| Minimum (ms) |
1.03 |
1.05 |
0.77 |
0.91 |
1.13 |
| Average (ms) |
1.24 |
1.16 |
0.79 |
0.94 |
1.26 |
| Maximum (ms) |
12.45 |
2.63 |
1.33 |
2.68 |
6.20 |
Memory (Read)
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Ops per Second |
3580840.03 |
820462.49 |
3780496.79 |
4495479.80 |
3316380.32 |
| Minimum (ms) |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
| Average (ms) |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
| Maximum (ms) |
2.85 |
0.10 |
1.21 |
1.18 |
5.97 |
Memory (Write)
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Ops per Second |
3606109.89 |
816121.10 |
3780543.62 |
4484131.67 |
3330445.50 |
| Minimum (ms) |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
| Average (ms) |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
| Maximum (ms) |
5.46 |
0.41 |
0.94 |
1.63 |
2.18 |
File I/O
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Reads per Second |
1292.30 |
1342.85 |
1816.20 |
3833.18 |
1987.88 |
| Writes per Second |
861.52 |
895.22 |
1210.78 |
2555.46 |
1325.23 |
| Fsyncs per Second |
2751.33 |
2856.59 |
3869.16 |
8170.37 |
4233.94 |
| Minimum (ms) |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
| Average (ms) |
0.26 |
0.20 |
0.14 |
0.07 |
0.13 |
| Maximum (ms) |
56.30 |
6.86 |
12.52 |
4.97 |
10.46 |
MySQL
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Transactions per Second |
1521.00 |
2354.33 |
3154.00 |
4230.33 |
2993.33 |
| Queries per Second |
30420.00 |
47086.67 |
63080.00 |
84606.67 |
59866.67 |
| Minimum (ms) |
3.64 |
2.30 |
2.29 |
1.64 |
2.04 |
| Average (ms) |
8.00 |
4.25 |
3.17 |
2.37 |
3.35 |
| Maximum (ms) |
112.40 |
47.12 |
17.80 |
19.98 |
55.58 |
Speed Test
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Distance (km) |
7877.22 |
7914.01 |
7914.01 |
7913.11 |
7916.01 |
| Latency (ms) |
128.095 |
125.635 |
127.272 |
117.742 |
122.570 |
| Download (Mbit/s) |
507.37 |
230.73 |
401.72 |
303.20 |
564.88 |
| Upload (Mbit/s) |
200.98 |
149.58 |
204.85 |
141.72 |
220.02 |
Apache Benchmark (against nginx on the servers)
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Requests per Second |
174.89 |
160.50 |
196.02 |
194.58 |
175.18 |
| Time per Request (ms) (mean) |
2911.50 |
3133.61 |
2570.81 |
2585.30 |
2858.36 |
| Transfer Rate (Kbyte/sec) |
146.71 |
134.63 |
164.43 |
163.23 |
146.95 |
Conclusion
Linode crushed the CPU benchmark this time out, and barely inched out on the
ab benchmarks.
UpCloud pulled ahead in nearly every other category with the exception of the
speed test. Latency was trivially better than the rest, but Vultr’s upload and
download speeds came in at the top.
Amazon Lightsail is still the top provider if you’re looking for the most
storage and transfer allotments.
Something quite interesting to note is that both Linode and Lightsail’s
distance’s came in exactly the same. Even though those numbers are an average,
that particular number usually reports the same distance for each of the servers
for a provider. I did verify that was the case this time. I’m guessing they are
sharing a data center.
