Welcome to WebmasterWorld Guest from 184.108.40.206
Forum Moderators: phranque
I'm trying to find a hosting company for my website.
Among other things, speed and reliability are very important to me.
My problem is, I'm finding it very hard to judge the speed and reliability of different hosting companies without first subscribing to the service.
When trying to search for an authoritative source on these matters all I am finding is a bunch of marketing hype and affiliate links.
Is there a good source that compares different hosting companies?
Also, if speed and reliability are important to me, should I not be looking at shared hosting? I've looked at dedicated and virtual dedicated servers, but again, all I'm finding is a bunch of hype. Are there benchmarks that they preform better?
Print magazines sometimes run such articles, but again, you still can see some bias creeping in. And then there's the case where a host has been doing well, and then there's a change in management and it all goes into the slop in a couple weeks.
This happened with a host I used for several of my sites -- a host that Inc magazine featured as one of the best and brightest severla years ago. Within 2 months (because of the article?) they were purchased by a bigger fish and never had the same level of service to this day.
if speed and reliability are important to me, should I not be looking at shared hosting?
Right, you should not be looking at shared hosting in my experience.
So it looks like just by eliminating shared hosting plans, that cuts out a lot of options I no longer need to consider.
That will help me when doing comparisons of different companies.
A question of clarification. Do virtual servers have the same speed and reliability as dedicated servers, or would they share the same pitfalls as shared hosting?
Also, if speed and reliability are important to me, should I not be looking at shared hosting? I've looked at dedicated and virtual dedicated servers
Yes, I would rule-out shared hosting.
Do not rule-out VPSs. But do make sure that you get a guaranteed CPU% (this will be the case with Xen VPSs, perhaps others), and make sure you know how many VPSs are hosted per physical machine, and what hardware is being used.
Virtual servers don't have the speed of dedicated servers, but perhaps you don't need a dedicated server. Also, make sure you compare apples with apples - A 1/10 share of a dual or quad high-end server could be as fast as 1/2 the speed of a low-end dedicated server (or more).
One advantage of virtual servers is that (given good management practices) you can be quickly moved from one machine to another in case of failure, need for added capacity, etc. Of course this should be true in theory with dedicated servers, as well. (The key being "good management practices".) Of course, if you own your own servers, you need to make your own redundancy and disaster-recovery plans.
Make sure you do the "phone test". Pick up the phone, and call the host, with a technical question. See how quickly you get to a real technician, and how well they are able to answer your question.
Consider the colocation facility (security, backup, etc.), it's physical and topographic location (is it in the network-topographic and physical center of your users?), the carriers (are they tier-1?) and the available carriers (can they expand?)
You want a host that connects to multiple tier-1 carriers in a major data center, not one at the end of a long one-lane road. traceroute to one of the host's customers (NOT the host's own website!) from multiple locations (you can find many websites around the world that will let you do this), and satisfy yourself that they have connectivity through multiple tier-1 carriers. Look for direct peering with major non-carrier networks as well - Cox, Time-Warner, AOL, Yahoo, Microsoft, etc. A good host will be in a datacenter that will peer directly with one or more of the above, who may even have servers on-site. (Do you need to access Google services, for example?)
It may not be easy to find hosts that meet your criteria, but it is easy to rule-out ones that don't. :)
I'm now down to a short list, and I'm willing to open an account on several to run some benchmark tests throughout a month to see how well the perform compared to each other.
Are there benchmark scripts or tools I can use to help me compare the different hosting accounts?