Featured Benchmarks DigitalOcean vs Linode vs Vultr [48GB 16 CPUS] centminmodbench.sh results

Compare DigitalOcean vs Linode vs Vultr

I ran comparison centminmodbench.sh benchmarks for 48GB memory 16 CPU processor core VPSes from 3 cloud / VPS hosting providers DigitalOcean, Linode and Vultr. Below are the tabulated results. May 6th, 2015 update - just a note that Vultr has increased their prices and changed their plans.

Referral links:

If you like this comparison review, please consider using my referral links to sign up with the below providers or put the word out and share this benchmark review with your friends etc :

• DigitalOcean - sign up via this link and you also get $10 credit to your account as well
• Linode - similar offer, use promo code LINODE10 for $10 credit to your account too
• Vultr - Vultr already offers free $5 credit for new sign ups

VPS servers:

For Vultr, New Jersey only location offering 16 cpu VPS

• DigitalOcean KVM VPS - San Francisco: 48GB, 16 CPUS, 480GB SSD 8TB (1Gbps network in/out) US$480/month
• Linode XEN VPS - Freemont California: 48GB, 16 CPUS, 1152GB SSD 20TB (40Gbps in & 6Gbps network out) US$480/month
• Vultr KVM VPS - New Jersey: 48GB, 16 CPUS, 600GB SSD 8TB (1Gbps network in/out) US$429.95/month

Testing Methodology

I basically created each respective VPS server using CentOS 6.5 64bit OS. Then ran the new installnbench.sh file at centminmod/centminmodbench · GitHub which I wrote specifically for this testing. It basically does 2 tasks when run installnbench.sh via curl command line on a fresh CentOS 6.5 or 7.0 OS server:

1. install latest Centmin Mod .08 beta01 LEMP stack ~ 15-30 mins time then
2. installs and runs centminmodbench.sh (UnixBench enabled) ~ 30-60 mins time

Code (Text):

curl -sL https://github.com/centminmod/centminmodbench/raw/master/installnbench.sh | bash

 

Centmin Mod .08 beta01 Install time

Centmin Mod .08 beta01 LEMP web stack install times are highly variable as part of the time is downloading source tarballs or YUM package downloads and installs so depends on the speed of the respective YUM mirrors the system picks up to download from as well as location of the VPS server geographically to the downloads. Downloads account for between 15-33% of the total install time. The rest is source compilation for Nginx, PHP-FPM, PHP extensions, Memcached server, CSF Firewall and other installed items via the Centmin Mod auto installer. Ultimately, comes down to cpu processor, disk performance and cpu clock speed for source compilation parts.

For installnbench.sh users, you'll find the install time on the very last line in the /root/centminlogs/centminmod_1.2.3-eva2000.08_***-***_install.log install log.

 

DigitalOcean vs Linode vs Vultr - ioping disk

Ioping measures random disk IO performance.

• For the default ioping test seems DigitalOcean > Vultr > Linode.
• For disk seek performance, Linode > DigitalOcean > Vultr
• For sequential disk ioping performance, DigitalOcean > Linode > Vultr
• For cached ioping disk performance, Vultr > DigitalOcean > Linode

 

DigitalOcean vs Linode vs Vultr - disk dd

Linode's SSD disk configuration is the winner here for disk dd benchmarks, followed by DigitalOcean and Vultr.

 

DigitalOcean vs Linode vs Vultr - disk fio

Vultr seems to have lower yet more uniform disk performance for random reads and random writes being roughly equal with 48K iops compared to DigitalOcean at 88K reads and 16K writes and Linode's 86K reads and 28K writes. However, both DigitalOcean and Linode have 83% and 79% respectively more random read IOPs than Vultr but lower random write IOPs.

 

DigitalOcean vs Linode vs Vultr - bandwidth tests

• Linode clearly has a better network in terms of bandwidth tests given that it's 48GB 16CPU VPS plan allows for respectively 40x and 6x times higher bandwidth network throughput with 40Gbps in and 6Gbps out compared to DigitalOcean and Vultr at 1Gbps in and out.
• Linode also has better performance for Asian test locations and just edging out DigitalOcean for European bandwidth locations.
• But note, Vultr is located on East coast USA in New Jersey, while DigitalOcean and Linode VPS are located on West coast USA.
• My personal preference due to visitor traffic for my sites is for West Coast USA as it has better latency and speeds to Asian/Australian locations.

Update: Some of the bandwidth speeds reported might be lower due to Centmin Mod .08 beta01 disabling tcp_sack. Re-enabling would allow most locations to gain in speed. So will be redoing comparison benchmarks for a part 2 when Vultr multi-core VPSes become available.

 

DigitalOcean vs Linode vs Vultr - ping benchmarks

• Ping tests are all geographically relative to the VPS server location, so West coast USA would fair better with same location ping test sites and vice versa for East coast USA based Vultr VPS.

 

DigitalOcean vs Linode vs Vultr - OpenSSL System Benchmarks

• OpenSSL system benchmarks test against system default OpenSSL 1.0.1e for number of ciphers and configurations would be good indicators for https SSL performance given Google now includes https in their ranking signal search engine algorithms. Relevant tests include rsa 2048bit, rsa 4096bit and 256bit ecdsa key algorithms. You can read why the newer ECC SSL certificates using 256 bit ecdsa key algorithms are better performing than the more commonly used rsa 2048 bit keys here. As well as test aes128/256 cbc non cpu hardware assisted performance and AES-NI cpu hardware assisted performance via the evp aes128/256 numbers.
• For DigitalOcean and Linode at least both are using Intel Xeon E5-2600v2 based processors based on Ivy Bridge EP platform which both support AES-NI (Advanced Encryption Standard New Instructions). So naturally you would see alot faster evp aes128/aes256 numbers compared to cpu processors not using AES-NI.
• Vultr just edges out Linode to take the top position for rsa 2048 bit, rsa 4096 bit and 256 bit ecdsa sign/s and verify/s. However, Vultr just falls behind Linode for AES-NI assisted evp aes256 and evp aes128 performance which may suggest Vultr might be using a early generation processor - maybe Intel Xeon E5-2600v1 Sandy Bridge EP based which still supports AES-NI but can't match the minor improvements in AES-NI performance that the newer Intel Xeon E5-2600v2 Ivy Bridge EP platform has. It could be Vultr using an Intel Xeon E5-2600v1 which is higher clocked compared to Linode's Intel Xeon E5-2680v2, which would explain the higher numbers for non-AES-NI based OpenSSL numbers for rsa 2048 bit, rsa 4096 bit and 256 bit ecdsa sign/s and verify/s. However, lscpu output shows Vultr has a lower base clock, so it could be the cpu used has a higher Turbo Boost clock speed in play ?

 

DigitalOcean vs Linode vs Vultr - Centmin Mod Nginx static OpenSSL Benchmarks

Centmin Mod LEMP web stack's Nginx install is compiled statically against a more up to date OpenSSL version 1.0.1i right now. This allows more flexibility in choosing what versions of OpenSSL to use for Nginx and Nginx SSL / SPDY SSL including using OpenSSL 1.0.2 with chacha20-poly1305 cipher support.

 

DigitalOcean vs Linode vs Vultr - mysqlslap MySQL Benchmarks

mysqlslap tool is included with all MySQL installs including Oracle MySQL, MariaDB MySQL and Percona MySQL. The custom mysqlslap parameters I used are just arbitrarily in nature and only test MyISAM performamce across 64 clients.

Actual test command used is:

MariaDB 10.0.13 MySQL is the default installed MySQL version in Centmin Mod .08 beta01 LEMP auto installer. Vultr comes out on top. Processor performance and clock speed plays a factor and probably Vultr leads because of this as well as maybe the more uniform and performant random write disk IOPs numbers shown above for disk fio benchmarks.

 

DigitalOcean vs Linode vs Vultr - PHP Benchmarks

• PHP benchmarks test the Centmin Mod .08 beta01 installed PHP 5.4.32 (php-fpm) binary (which defaults include enabling APC Cache 3.1.13 with igbinary support) and tests against the PHP 5.4.32 source included Zend/bench.php and Zend/micro_bench.php files.
• You can see the PHP 5.4.32 default compiled extensions here.
• 3 runs are made and average taken for time it takes to run the benchmarks.
• This test is single threaded in nature so doesn't scale with multiple cpu cores/threads and the underlying performance is based partly on cpu processor performance and clock speed. Which the final results confirm as Vultr is the winner, then Linode and DigitalOcean with the lower clock speeds comes out last.

PHP 5.4.32 default configuration

PHP 5.4.32 default loaded modules

Code (Text):

[PHP Modules]
apc
bcmath
bz2
calendar
Core
ctype
curl
date
dom
enchant
ereg
exif
filter
ftp
gd
gettext
gmp
hash
iconv
igbinary
imagick
imap
intl
json
libxml
mbstring
mcrypt
memcache
memcached
mhash
mysql
mysqli
mysqlnd
openssl
pcntl
pcre
PDO
pdo_mysql
pdo_sqlite
Phar
posix
pspell
readline
Reflection
session
shmop
SimpleXML
soap
sockets
SPL
sqlite3
standard
sysvmsg
sysvsem
sysvshm
tidy
tokenizer
xml
xmlreader
xmlrpc
xmlwriter
zip
zlib

 

DigitalOcean vs Linode vs Vultr - UnixBench Benchmarks

As per below notes, you can only validly compare UnixBench system index scores within the same branch of Linux kernels 2.6.x or 3.x. Comparing 2.6.x vs 3.x kernels won't be directly comparable due to differences in subtest numbers that the index score is derived from.

Vultr and DigitalOcean use same 2.6.x linux kernel, so you can see definitely that Vultr is a faster system than DigitalOcean. To compare with Linode, you can probably only rely on looking at the Dhrystone and Whetstone numbers to gauge cpu processor performance.

• DigitalOcean uses Xeon E5-2630L processors with 6 cores / 12 threads clocked at 2.4Ghz with Turbo Boost to 2.8Ghz. All cores Turbo Boosts to 2.6Ghz and single cpu core Turbo Boosts to 2.8Ghz
• Linode uses Xeon E5-2680v2 processors with 10 cores / 20 threads clocked at 2.8Ghz with Turbo Boost to 3.6Ghz. All cores Turbo Boosts to 3.1Ghz and single cpu core Turbo Boosts to 3.6Ghz.
• Vultr was not able to determine which processors used from benchmark numbers but it looks to be faster than Linode's Xeon E5-2680v2 given the UnixBench subtest numbers from Dhrystone, Whestone and above cpu hardware assisted AES-NI OpenSSL evp aes128/aes256 benchmark numbers. Could be that Vultr is using a higher clocked Xeon E5-2600v2 most likely ? Xeon E5-2690v2 maybe ? The lscpu numbers show Vultr at 2.6Ghz base clock so it could be Xeon E5-2650v2 or an Xeon E5-2670v1 ?
• Looks like based on Dhrystone, Whestone and OpenSSL tests, Vultr > Linode > DigitalOcean.

Intel ARK comparison of Xeon E5-2630L vs Xeon E5-2680v2

Notes regarding UnixBench Index numbers

Unixbench index numbers are much lower for Linux 3.x kernels as opposed to Linux 2.6.x kernels as some sub tests are influenced by the differing Linux Kernel versions used. From previous local tests I have done with CentOS, the 2.6.32 Kernels performed better in some UnixBench subtests than 3.xx Kernels which are used by Linode. Particularly, for Process Creation Index tests.

Code (Text):

2.6.32-431 = 3287.4
3.8.13-26.1 = 2417.4
3.13.4-1 = 1341.0

You can read the explanation at byte-unixbench - A Unix benchmark suite - Google Project Hosting which suggests subtests like pipe throughout has no real world bearing on performance so can be disregarded

 

DigitalOcean vs Linode vs Vultr Summary

The reason why I did the above comparison benchmarks is because I use all 3 VPS cloud hosting providers to constantly test my Centmin Mod LEMP menu based auto installer web stack as well as having this very forum hosted on Linode right now. So it's good to know which VPS cloud hosting provider has the best performance for my needs which includes Centmin Mod install times (I may install and re-install 100s of times per month). As well as general performance that has some relative real world applications for what a normal LEMP web stack would be for - Nginx SPDY/3.1 SSL, PHP-FPM performance and MariaDB MySQL performance.

Given the monthly prices are pretty much the same with Vultr slightly cheaper, it seems Linode is the better value as it has more SSD disk space at 1152GB vs 480GB for DigitalOcean and 600GB for Vultr, more bandwidth allocated 20TB vs 8TB, and faster network in/out at 40Gbps/6Gbps vs 1Gbps.

For Centmin Mod install times:

• Vultr is clearly faster for the above tests at 880 seconds to install the entire Centmin Mod .08 beta01 LEMP web stack - just under 15 minutes. I had previously benchmarked lower 2-8 CPU core DigitalOcean and Linode VPSes with an earlier Centmin Mod .07 beta and there Linode won over DigitalOcean.

For Disk Performance:

• There's no clear leader. However, overall Linode seems to have the right mix of sequential and random disk IO read/write performance. Followed by DigitalOcean and the Vultr.

For Bandwidth & Pings performance:

• Linode has the best network configuration with 40Gbps in and 6Gbps out for their 48GB 16CPU VPS plan. It shows in some of the numbers. However, you do need to note that Vultr was tested on the East coast USA in New Jersey so better suited for Europe regions as well while the other two were tested on West Coast USA which would be better suited to Asia/Oceania regions.

For OpenSSL performance:

• Nginx SPDY SSL and https performance is more important now given Google's announcement. So if you are planning to use Nginx SPDY/3.1 SSL based web server for your sites and forums, you would want to pick a VPS provider with the better Intel Xeon E5 based processors. Linode has the best AES-NI assisted aes128/aes256 performance while only just behind Vultr for non-AES-NI performance.
• For SSL certificates related usage, the most commonly used key algorithms will be either rsa 2048 bit, rsa 4096 bit and 256 bit ecdsa. With that in mind, Vultr will have the edge over Linode and both Vultr and Linode would perform better than DigitalOcean probably due to the lower cpu processor clock speed used on the Intel Xeon E5-2630L processors.

For mysqlslap MySQL performance:

• Vultr > Linode > DigitalOcean. I suspect partly due to cpu clock speeds although unsure of what processor Vultr uses. But MySQL performance tested was for MyISAM engine and default my.cnf tuned by Centmin Mod installer. So doesn't really reflect real world usage and when you use InnoDB engine or when you properly tune your MySQL my.cnf settings for your work loads.

For PHP performance

• The PHP benchmarks were only single threaded so reflects more cpu processor performance and clock speed. Again Vultr > Linode > DigitalOcean.

For UnixBench performance:

• Again, Vultr > Linode > DigitalOcean due to probably cpu processor clock speed factors.
• Below are the lscpu outputs for all 3 VPS servers. But Vultr based cpu clock is lower than Linode according to lscpu output ? So maybe Vultr cpu has higher Turbo Boost clock speeds than Linode when not all cpu threads are utilised and engaged ? This would show itself in tests which do not use all available cpu cores/threads. Or could be difference in overhead for Linode Xen based virtualization vs Vultr KVM based virtualization ?

DigitalOcean

Code (Text):

----------------------------------------------
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                16
On-line CPU(s) list:   0-15
Thread(s) per core:    1
Core(s) per socket:    1
Socket(s):             16
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 62
Stepping:              4
CPU MHz:               2399.998
BogoMIPS:              4799.99
Virtualization:        VT-x
Hypervisor vendor:     KVM
Virtualization type:   full
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              15360K
NUMA node0 CPU(s):     0-15

Linode

Code (Text):

----------------------------------------------
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                16
On-line CPU(s) list:   0-15
Thread(s) per core:    16
Core(s) per socket:    1
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 62
Stepping:              4
CPU MHz:               2800.046
BogoMIPS:              5602.42
Hypervisor vendor:     Xen
Virtualization type:   full
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              25600K
NUMA node0 CPU(s):     0-15

Vultr

Code (Text):

----------------------------------------------
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                16
On-line CPU(s) list:   0-15
Thread(s) per core:    1
Core(s) per socket:    16
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 42
Stepping:              1
CPU MHz:               2599.998
BogoMIPS:              5199.99
Hypervisor vendor:     KVM
Virtualization type:   full
L1d cache:             32K
L1i cache:             32K
L2 cache:              4096K
NUMA node0 CPU(s):     0-15

For Vultr cpu, with SMP capable Intel Xeon cpus with a base clock of 2.6Ghz, narrowed it down to maybe of these according to Intel ARK database ? Still waiting on reply from Vultr as to specific model used.

• Intel Xeon Processor E5-2670 (20M Cache, 2.60 GHz)
• Intel Xeon Processor E5-4650L (20M Cache, 2.60 GHz)
• Intel Xeon Processor E5-2630 v2 (15M Cache, 2.60 GHz)
• Intel Xeon Processor E5-2650 v2 (20M Cache, 2.60 GHz)
• Intel Xeon Processor E5-4607 v2 (15M Cache, 2.60 GHz)
• Intel Xeon Processor E5-4620 v2 (20M Cache, 2.60 GHz)
• Intel Xeon Processor E7-4860 v2 (30M Cache, 2.60 GHz)

 

You're welcome - time and $$$ hehe. I'll be testing the higher cpu plans from DigitalOcean 20 cpu cores, Linode 20 cpu cores and Vultr 24 cpu cores later. Unfortunately, Vultr doesn't have 24 cpu core plans available right now

 

haha don't have that kind of $$$$ hehe

 

maybe start a new thread for centminmodbench.sh for dedicated servers in dedicated server forums