VPS comparison OVH Vs Amazon ec2 Vs Google cloud
I’m going to migrate busylog.net to a VPS hosting. At the moment busylog.net is hosted on OVH shared hosting. This mean that I share IP and resources with other many sites and, more annoying, I can’t access to the server.
In order to have direct access to server (through SSH) and to have more control on server (execute root commands, add PHP and Apache modules, add services…) I’m going to migrate to a VPS.
Target is to spent not more than ~$5/mo.
I need very few hardware resources.
Below my notes and some benchmark considering 3 VPS providers:
As first overview below in the Tabbed parts of this notes.
Conclusion (start at the end after 5 months)
Starting from tab:“|OVH Vs. Amazon Vs. Google|” you can find some notes about my comparison among VPS. Instead about my conclusion I chose (~5 months ago) to go with OVH VPS SSD (here). After 5 months I can confirm that OVH offer is very attractive in terms of price, features, usability and service availability.
1 year I spent 42$ (51$ with tax) : 3.5$/months.
Moreover costs are fixed … instead with Amazon and Google according to me is a little more tricky figure out the final cost.
OVH bandwidth is throttled (outbound and inbound) to ~10MB/s (~100Mbps) that is anyway enough for me . With OVH there are no throttling (like Amazon) nor closure (like Google) on TCP ports. You can have for free 1 extra IP (i.e. if you need for SEO reasons) located in some region. About CPU/Disk… really this doesn’t matter for my use purposes however with Google and Amazon you can’t have 2GB/RAM and 10GB/Disk for this price.
OVH Manager Usability
The OVH VPS offer is for an “not managed” hosting: for example if you get a new IP you have to configure the linux-box (adding IP in this example) yourself: your are alone.
However the web portal provides some actions such as:
install VPS (with several Linux distribution), reset password (rebooting in “Rescue” mode), reboot VPS, KVM (web shell console), manage extra disks, Backups and SnapShot …
Benchmark results depends on several factors (of course as all now) one of this is geolocation of servers (geolocation of load sources respect to target).
Disclaimer To be honest I didn’t put enough effort to perform a proper benchmark test as I’m used to doing for my tasks. Despite you will see numbers, load graphics and video comparison … the correct meaning is in terms of qualitative comparison (X this is probably better than Y).
About geolocation I chose a site geographically near to most parts of visitors of busylog.net (OVH datacenter in Beauharnois).
OVH Shared Hosting is instead placed in Europe. As load source (for the comparison benchmark) I chose a source geographically near to visitors of site (and consequently close to VPS) .Disclaimer2 The below comparison is OVH VPS Vs. OVH Shared Hosting. About VPS I installed out-of-the-box without any kind of SO, MySQL and wordpress optimization (I just copied the WordPress directory from OVH Shared Hosting and copied to OVH VPS).
⭕ Response time
↪ webpagetest: OVH Shared Hosting Vs. VPS
(video comparison: load time)
↪ loadimpact: OVH Shared Hosting
↪ loadimpact: OVH VPS
↪ Google: OVH Shared Hosting
↪ Google: OVH VPS
↪ pingdom: OVH Shared Hosting
↪ pingdom: OVH VPS
⭕ Is site rank increased?
After the site is migrated from OVH Shared Hosting to VPS the daily number of pages view is increased by +150%. I guess this increasing of pages view is due to site optimisation that has decreased the load time (~1sec) and increased the page score in PageSpeed Insights. Moreover (probably) the rank on google takes advantages by the IP geolocated in USA and not shared with other hosted sites.
Enter a URL to test the load time of that page, analyze it and find bottlenecks.https://www.webpagetest.org/
⭕Script: busytest.sh Details here
⭕Script: bench.sh Details here
⭕Script: vhwinfo.sh Details here
⭕Script: speedtest.py Details here
2 GB RAM
SSD 10 GB
100 Mbps (12MB/s)
0.5 GiB RAM
HD/SSD 10 GB
Pay extra for traffic
Order needs +3 days to be completed
All network ports are open
Extra IPs (*ip_notes ) only ~$2 one-time ($0/mo)
No extra costs
~60 sec. for a new instance
TCP Port 25 is throttled
~60 sec. for a new instance
TCP Port 25 is closed
In EU only for business
Quebec,Canada North America …
We can buy (for $2) one more IP located in US:
United States New Jersey …
Please consider that GEO IP location is based on DB information of geo-coordinates. So several times an IP that you bought from OVH for US is located as results in Canada.
I/O (1st run): 476 MB/s
I/O (2nd run): 432 MB/s
I/O (3rd run): 437 MB/s
Average I/O : 448.3 MB/s
I/O (1st run): 19.7 MB/s
I/O (2nd run): 21.4 MB/s
I/O (3rd run): 21.3 MB/s
Average I/O : 20.8 MB/s
General Purpose SSD
I/O (1st run) : 67.9 MB/s
I/O (2nd run) : 64.3 MB/s
I/O (3rd run) : 64.2 MB/s
Average I/O : 65.4667 MB/s
I/O (1st run): 37.4 MB/s
I/O (2nd run): 37.5 MB/s
I/O (3rd run): 37.5 MB/s
Average I/O : 37.46 MB/s
SSD ( see notes_SSD )
I/O (1st run) : 37.4 MB/s
I/O (2nd run) : 37.6 MB/s
I/O (3rd run) : 35.5 MB/s
Average I/O : 36.8333 MB/s
San Jose 11.2MB/s
(*) note OVH limits bandwidth to 100 Mbps (12MB/s)
San Jose 23.5MB/s
San Jose 16.0MB/s
(please note Mbit/s
Montreal, QC [31.08 km] ping : 2.765 ms
Dow.d: 99.86 Mbit/s
Upl.d: 99.16 Mbit/s
Pendleton, OR [73.21 km] ping : 19.563 ms
Dow.d: 282.66 Mbit/s
Upl.d: 214.94 Mbit/s
Naples [9345.30 km]
why? HD 9.8/8.1 GB -> Average I/O : 37.46 MB/s SSD -> Average I/O : 36.8333 MB/s ? [Google-busylog_net@instance-SSD ~]$ df -h Filesystem Size Used Avail Use% Mounted on /dev/sda1 9.8G 1.3G 8.1G 14% / tmpfs 296M 0 296M 0% /dev/shm [Google-busylog_net@instance-SSD ~]$ cat /sys/block/sda/queue/rotational 0 SSD (Virtually) [Google-busylog_net@instance-SSD ~]$ sudo hdparm -W0 /dev/sda /dev/sda: setting drive write-caching to 0 (off) SG_IO: bad/missing sense data, sb: 70 00 05 00 00 00 00 0a 00 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 SG_IO: bad/missing sense data, sb: 70 00 05 00 00 00 00 0a 00 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 SG_IO: bad/missing sense data, sb: 70 00 05 00 00 00 00 0a 00 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 SG_IO: bad/missing sense data, sb: 70 00 05 00 00 00 00 0a 00 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 SG_IO: bad/missing sense data, sb: 70 00 05 00 00 00 00 0a 00 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 write-caching = not supported [Google-busylog_net@instance-SSD ~]$ dd if=/dev/zero of=/tmp/testfile bs=100M count=1 oflag=direct 1+0 records in 1+0 records out 104857600 bytes (105 MB) copied, 2.8136 s, 37.3 MB/s --- TODO (complete notes) / sorry ---
Specific OVH tests
⭕Test OVH speed outbound (~10MB/s ports: 80 and 22)
Google 80: OVH #google_server> wget http://<OVH_SERVER>/FILES2DW/100MB.zip 104,857,600 11.9M/s in 9.5s 2016-12-13 20:35:15 (10.5 MB/s) - "100MB.zip" saved [104857600/104857600] Google 22: OVH #google_server> scp <OVH_SERVER>:/var/www/html/FILES2DW/1GB.zip . 1GB.zip 100% 1024MB 11.4MB/s 01:30 Amazon 80: OVH #Amazon_server> wget http://<OVH_SERVER>/FILES2DW/100MB.zip 100.00M 9.80MB/s in 11s 2016-12-13 20:56:36 (8.72 MB/s) - '100MB.zip' saved [104857600/104857600] Amazon 22: OVH #Amazon_server> scp root@<OVH_SERVER>:/var/www/html/FILES2DW/1GB.zip . 1GB.zip 100% 1024MB 9.4MB/s 01:49
Oregon, United State