By request, we’re revisiting the Singapore data centers for this month’s VPS
Showdown.
Incidentally, while not part of the benchmarks currently, Linode did expand
their S3-compatible Object Storage to the Singapore data center. They also
turned 17 in mid-June with their usual slew of announcements, and upgrades for
existing customers.
Sadly, and as expected, AWS is still lagging behind the pack in terms of support
for Ubuntu 20.04 LTS, so this month’s benchmarks are still on Ubuntu 18.04 LTS.
Per the usual, I’ve spun up three instances with each provider and averaged the
results.
Overview
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Location |
Singapore 1 |
ap-southeast-1 |
Singapore, SG |
SG-SIN1 |
Singapore |
| RAM |
1 GB |
1 GB |
1 GB |
1 GB |
1 GB |
| CPU |
1 Core |
1 Core |
1 Core |
1 Core |
1 Core |
| Storage |
25 GB SSD |
40 GB SSD |
25 GB SSD |
25 GB SSD |
32 GB NVMe SSD |
| Transfer |
1 TB |
2 TB |
1 TB |
1 TB |
1 TB |
| Base Price |
$5/month |
$5/month |
$5/month |
$5/month |
$6/month |
| Backups |
$1/month |
N/A |
$2/month |
$2/month |
$1.20/month |
| Transfer Overage |
$0.01/GB |
$0.09/GB |
$0.01/GB |
$0.01/GB |
$0.01/GB |
| Load Balancer |
$10/month |
$18/month |
$10/month |
N/A |
$10/month |
| Block Storage |
$0.10/GB |
$0.10/GB |
$0.10/GB |
$0.223/GB |
$0.10/GB |
| Object Storage |
Yes |
Yes |
Yes |
No |
Yes |
| Managed Databases |
Yes |
Yes |
On 2020 Roadmap |
No |
No |
| 2FA/MFA |
Yes |
Yes |
Yes |
Yes |
Yes |
| One-click Apps |
Yes |
Yes |
Yes |
No |
Yes |
| Custom Images |
Yes |
No |
Yes |
Yes |
Yes |
CPU Info
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| CPU MHz |
2200.00 |
2398.21 |
2066.66 |
2995.31 |
3792.00 |
| Cache Size (KB) |
4096.00 |
30720.00 |
512.00 |
16384.00 |
16384.00 |
| BogoMips |
2933.00 |
3200.33 |
2666.67 |
3995.00 |
5056.33 |
CPU
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Events per Second |
772.51 |
865.64 |
1229.91 |
1124.54 |
1166.42 |
| Minimum (ms) |
1.17 |
1.06 |
0.75 |
0.82 |
0.74 |
| Average (ms) |
1.29 |
1.15 |
0.81 |
0.90 |
0.87 |
| Maximum (ms) |
4.34 |
2.67 |
7.71 |
3.06 |
3.26 |
Memory (Read)
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Ops per Second |
3119758.47 |
832212.29 |
3423331.05 |
4342445.14 |
4677596.32 |
| 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) |
0.97 |
0.44 |
1.52 |
0.39 |
1.97 |
Memory (Write)
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Ops per Second |
3166084.42 |
828554.36 |
3444632.03 |
4359066.80 |
4755396.89 |
| 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) |
0.63 |
2.35 |
1.12 |
0.85 |
1.57 |
File I/O
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Reads per Second |
993.53 |
1494.47 |
1609.36 |
2724.36 |
3542.82 |
| Writes per Second |
662.34 |
996.29 |
1072.90 |
1816.23 |
2361.85 |
| Fsyncs per Second |
2115.01 |
3181.13 |
3421.40 |
5803.29 |
7550.80 |
| Minimum (ms) |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
| Average (ms) |
0.27 |
0.17 |
0.16 |
0.10 |
0.08 |
| Maximum (ms) |
484.32 |
8.94 |
100.93 |
12.55 |
8.94 |
MySQL
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Transactions per Second |
1370.33 |
2367.33 |
2320.67 |
3687.67 |
4487.00 |
| Queries per Second |
27406.67 |
47346.67 |
46413.33 |
73753.33 |
89740.00 |
| Minimum (ms) |
4.10 |
2.35 |
2.66 |
1.60 |
1.45 |
| Average (ms) |
7.96 |
4.23 |
4.32 |
2.91 |
2.39 |
| Maximum (ms) |
990.31 |
29.77 |
18.36 |
26.93 |
18.02 |
Redis
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| PING_INLINE |
36670.35 |
56333.11 |
60130.16 |
55415.64 |
48111.17 |
| PING_BULK |
36273.71 |
55564.35 |
57748.04 |
53987.68 |
47719.40 |
| SET |
36106.62 |
56215.92 |
60939.24 |
54888.28 |
47413.16 |
| GET |
36089.19 |
56015.41 |
61141.29 |
54969.82 |
43797.94 |
| INCR |
36493.64 |
56532.38 |
61039.45 |
55985.21 |
45878.18 |
| LPUSH |
37443.91 |
55679.39 |
56574.93 |
56900.00 |
49110.07 |
| RPUSH |
36823.02 |
56563.12 |
57881.36 |
55565.57 |
47250.34 |
| LPOP |
36674.72 |
55387.85 |
55593.88 |
55987.53 |
47633.78 |
| RPOP |
36881.14 |
55533.89 |
60123.99 |
54917.53 |
47458.54 |
| SADD |
36685.26 |
56861.95 |
61361.03 |
54723.85 |
46555.66 |
| HSET |
37077.44 |
55401.97 |
57818.72 |
55939.35 |
48994.33 |
| SPOP |
35732.44 |
56508.86 |
57622.25 |
56861.04 |
47517.98 |
| LRANGE_100 (first 100 elements) |
20929.18 |
30097.23 |
23078.91 |
29268.74 |
27749.95 |
| LRANGE_300 (first 300 elements) |
8376.19 |
12085.97 |
8260.53 |
12786.10 |
12212.72 |
| LRANGE_500 (first 500 elements) |
6041.12 |
8445.55 |
5979.53 |
8941.63 |
8900.24 |
| LRANGE_600 (first 600 elements) |
4995.95 |
6709.91 |
4676.54 |
7021.85 |
7232.30 |
| MSET (10 keys) |
29803.45 |
40895.60 |
30144.91 |
49184.35 |
41703.80 |
Speed Test
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Distance (km) |
10.71 |
6.13 |
6.13 |
9268.67 |
1777.07 |
| Latency (ms) |
2.159 |
2.287 |
8.315 |
3.940 |
2.269 |
| Download (Mbit/s) |
3263.73 |
908.13 |
3046.09 |
952.50 |
1649.36 |
| Upload (Mbit/s) |
1970.16 |
909.85 |
2149.07 |
1044.15 |
2584.66 |
Apache Benchmark (against nginx on the servers)
| |
DigitalOcean |
Lightsail |
Linode |
UpCloud |
Vultr |
| Requests per Second |
218.03 |
250.07 |
728.84 |
603.62 |
201.41 |
| Time per Request (ms) (mean) |
2500.75 |
2011.39 |
1376.97 |
1406.39 |
2552.27 |
| Transfer Rate (Kbyte/sec) |
182.70 |
209.78 |
611.41 |
506.36 |
168.96 |
Conclusion
While the Vultr High Frequency is still performing quite well for being only a
buck more, it seems like Linode’s focus on fleet stability and performance is
starting to show as they have began creeping up in the metrics.
DigitalOcean came out quite strong in terms of download speed with Vultr taking
the crown for uploads. While not being the fastest in either category
explicitly, their overall speed was quite impressive.
The distance for UpCloud and Vultr both seemed a bit off with this month’s speed
test, and I have to assume that’s due to the IP addresses allocated to those
instances may not have the correct geographical information associated with
them. Thinking maybe it’s best to drop that metric since it really doesn’t have
much bearing on things.
The thing to stand out to me this month is how exceptionally high DigitalOcean’s
maximum time metrics were. I’d much rather have a provider that has slower
overall metrics as long as the range of times is tighter.
