VPS vs. Share Account Performance (Apples vs. Oranges)
In the post earlier today, I said 1and1.com has their servers heavily loaded. At the same time, I was curious to what the performance level of my new VPS was when virtually idle. Here are some screen shots of the server load at 1and1 and my VPS as KnownHost.com using PHPSysInfo.
KnownHost.com VPS – with little to no load
1and1.com – Shared Account
Sure this is not apples to apples comparison and the boxes are very different in capacity. The VPS has dedicated ram and processing time. So I am not suggestion, that it is a fair comparison. It is interesting to see that 1and1.com does max out a server, which is reasonably typical in shared account hosting.
I do find it interesting the different CPUs and performance levels (BogoMips)
Below is the system specs for the two servers…
KnownHost.com
1and1.com
Long and short of it, shared accounts are fine for low use and cheap hosting. VPS/Dedicated or even a performance based shared hosting account is for the domain(s) that get a heavier use. Simply, you get what you pay for. I just went from $5.00 a month to $35 a month, that should give me a significant level of performance improvement.
Update: As I have continued getting used to the VPS environment and testing the server, I have just been floored by how heavily used the server at 1and1.com was/is. I have a been using a Vista Gadget called “Server Load” to watch the VPS. I loaded it on the 1and1.com shared account and “wow” is all I can say…. The screen shots for the server load of both the VPS and 1and1.com.
So the server load at 1and1.com has ranged from 4.x to 15.x. I did see it once at 35.x, but that only lasted for a few minutes. According to wikipedia load average is “a measure of the amount of work that a computer system performs”. So with the 11.4 in the screen shot, that equates to 1140% over capacity. Or the computer would need to be 11 times faster to keep up with demand. Do not this is not CPU utilization, but how long it takes for a command to be processed.
Simply, 1and1.com packs the box to maximize profits. Which in of itself is not a bad thing, just not a good thing when you require more resources. Remember the 1and1.com server is a single core Xenon and the VPS is dual-core, which should help the VPS be more efficient.
Related posts:
- The move to the new server is all done (mostly)
- Move to VPS Done (aka We’re Home for Thanksgiving)
- Scale
- Twitter Weekly Updates for 2010-01-03
- Review: Cooler Master Hyper N520
[...] http://machinadei.com/blog/2009/11/0…es-vs-oranges/ [...]
[...] http://machinadei.com/blog/2009/11/0…es-vs-oranges/ [...]
[...] http://machinadei.com/blog/2009/11/0…es-vs-oranges/ [...]
The “server load” metric comparison you make is not fair, because in a virtualized VPS environment the “server load” is not the real server load but a virtualized one. And, because it is virtualized, it could mean almost anything, i.e. the physical server could be on its knees and your VPS server load still show at almos 0 if your VPS was doing little.
@Pepe
Thanks for the comment…. and I do not claim this is an equal assessment, especially since it was an idle website (VPS) verses a more heavily used site (shared host). I did say it was two very different environments, one with more dedicated resources and one without any. It is why I say it is an “apples to oranges” comparison. They both may be fruit, but that is where the similarities end.
So sure the server my VPS could very well be on its knees, but at least I would still have some resources left (at least in accordance to my contract) and there is a basic performance level that is available in a VPS environment that is not in a shared host. Also you get what you pay for… $5 a month for unlimited everything, is a lot different than $35 plus for a VPS with lots of limits and constraints. Assuming the host is honest, which is what most users say about KnownHost.com, the server should not be on its knees and they should be managing use based on the hosts on each server.
All this was for is that I find it interesting the performance difference between one host and another. Outside of that, make what you will of it.
[...] of the move, I started looking into any issues of moving from the shared hosting environment and the VPS. The first issue is to modify the “wp-config.php” file to change DB information. From there you [...]