Why do hosts lie? Do they not realize users can usually check the info they give? Or do they just get so many customers that don't check, they don't care if a few catch them in a lie. I believe the latter.

Here's an example. I hosted with internetconnection.com (aka internetconnection.net). I asked their programmer(not just Customer Service) over the phone, "How many people are on the same server with me". They stated 200. From telnet, I counted over 500 user names on the server, so that was not true.

Also, I had a nightmare with their billing; and support wasn't great either. They do appear to own their own NOC. The owner, Brent, is really nice though. This was a few months ago. I hope they've improved.

Originally posted by chrisb
They do appear to own their own NOC.
Just out of interest, what do you think the % is of hosts who do not own their own datacentres?

Maybe they are resellers who lease a % of a server or maybe they lease managed dedicated server/s, or maybe they colo their own boxes, or maybe they manage their own server at a place like rackshack?? Few hosts actually own their own datacentres these days.

I would rather be on a box from a client of rackspace than a host who has scrimped together their own datacentre. You would most likely get better performance from a client of rackspace than the host that has invested capital into their own infrastructure. I'm very doubtfull that they could compare to the level of investment made by a company like rackspace.

Therefore if you host with a client of rackspace, you enjoy the benefits of the massive investment into infrastructure that rackspace has made. If you host with a host that has their own very small datacenter, then you run the risk of untold downtime as their cashflows are possibly stretched due to the investment of capital they have made into their own equipment.

The servers your site is hosted on and the surronding infrastructure would be nowhere near as advanced as a company like rackspace. So you are in most cases much better off hosting with a hosting provider that has a supplier like rackspace, than a hosting company with their own datacentre, IMHO. :)

This post was a bit longer than I intended. :)

I'd guess that less than 5% of hosts own their own data centers.

Do you think that he may have meant "web sites" and not "users"? If someone asked me how many users do I put per server I would take that to mean web sites, I have only marginal control over how many "users" are created, since each web site can create multiple users.

Originally posted by Dixiesys
Do you think that he may have meant "web sites" and not "users"?

Huh? I don't know why you'd think that. It seems pretty clearly stated to me.

Also, there were other incidents. That was only one example. Further, they never could get IMP webmail to work right either. I prefer Neomail anyway.

Hi Gary,
I agree with your point about a company similar to rackspace being better than a small struggling data center.

What I meant by that was at least they weren't a reseller OR a reseller of a reseller of a reseller. Though that's not bad either, if you get better service at the same price or lower.

Originally posted by chrisb


Huh? I don't know why you'd think that. It seems pretty clearly stated to me.

Also, there were other incidents. That was only one example. Further, they never could get IMP webmail to work right either. I prefer Neomail anyway.

Because if you called me on the phone tomorrow and asked me this same question, I'd say "around 225-250".

[root sites]# ls | grep www | wc -l
227
[root sites]# cat /etc/passwd | wc -l
652

227 web sites, 652 usernames and the 652 could easily go up another couple hundred as those web sites add a few email addresses. Each site on average was allotted 25 email addresses (which when added are users in the system).

I'm not defending anyone just saying perhaps you were simply talking apples and being told oranges.

Originally posted by chrisb
Hi Gary,
Errrr, I'm Aussie Bob :eek: :eek2:
I agree with your point about a company similar to rackspace being better than a small struggling data center.
:)
What I meant by that was at least they weren't a reseller
Ahhh, there are some pretty big hosts out there and in this forum who are "resellers".
OR a reseller of a reseller of a reseller. Though that's not bad either, if you get better service at the same price or lower.
Suffice to say, the general public has no idea how the hosting industry works. They think that all hosts have their own 50 MILLION dollar datacenter and 20 techicians working in white overalls. :rolleyes: :D

What would you consider a datacentre though? People are starting up "datacentres" in their garage's, buying a 10mbit line from cogent and starting from their.

Originally posted by Myacen
What would you consider a datacentre though? People are starting up "datacentres" in their garage's, buying a 10mbit line from cogent and starting from their.
Good point. I guess we all have a description of a "datacentre" in our own mind and there is not 1 that is the right description etc. What would you consider a datacentre anyways?

What about my 14.4 kbps Dial up line connected to my 386? *grinning* :)

Personally, datacenters can mean a lot of things, from a couple of servers hooked up to a router and a T1 line, to a Rackspace/Rackshack/IBM sized monster with its own OC-768s.

Think about it. If you are only hosting a few small sites, a T1 is fine. If you are trying to become an enterprise host, though, then you need something much bigger. :)

- James

I'd be quite happy with 3 T-1s giving me 4.5 megabits into a nice 12x12 room within a nice building someplace and a couple of racks and maybe some shelves for tower cases, a natural gas generator outside for backup power and a big ups inside to make the transition smooth. A good window unit air conditioner could serve as backup cooling. Of course I'd want bars on any windows and one really beefy door with a massive lock!

As I understand it you can tie up to 6 T-1s together for 9 megabits, that's not a shabby amount of bandwidth for a smaller datacenter. The hardest part would be trying to generate some true network redundancy in the local loop. Once you get to your provider they'll surely be multi-homed, but that doesn't help much if the local loop goes screwy between them and you.

It depends on what they offer prices and features. I would really like to host somewhere that has their own...but can you find one that does fit your needs and so on...

I consider a room/garage/office space filled with servers to be a "server room"

A dedicated building for a "server room" would be a "Data Center"

That's just how I see it though.

"How many people are on the same server with me"

If you asked me this question, I would tell you how many websites there are on the server. User names? No way.

Do you think that he may have meant "web sites" and not "users"?

Huh? I don't know why you'd think that. It seems pretty clearly stated to me.

Uhh... you said "How many people" which is not terribly clear. Most people don't care how many usernames, just how many websites. So unless you say "How many usernames, not websites, but usernames do you have on this server" you're going to get the same answer you got. :rolleyes:

Originally posted by Lurleene

Uhh... you said "How many people" which is not terribly clear. Most people don't care how many usernames, just how many websites. So unless you say "How many usernames, not websites, but usernames do you have on this server" you're going to get the same answer you got. :rolleyes:

Maybe I was unclear. I think I did ask how many "users". So, it's better to ask "How many usernames" OR "How many top level domains? [are on my shared server]" I just did a cd to the usr directory where I counted all the usernames. That's how I came up with over 500 users.

I'm sorry, but I agree with the others. The reason is because the question we hear all the time is "how many sites are hosted on the server"? And, although number of sites isn't very meaningful itself without knowing the nature of those sites, number of users would seem to have even less value.

Of course, if I get the question now, I will either ask for a clarification or include my assumption of the question in my response.

Originally posted by chrisb
I asked their programmer(not just Customer Service) over the phone, "How many people are on the same server with me". They stated 200. From telnet, I counted over 500 user names on the server, so that was not true.
This sounds like a case of the programmer "assume"ing they knew what the question was. We all do it, but it can still muck things up -- if not clarified; which could have be done in the asking of the question as well.

This could have turned out ever worse (funnier?) if the programmer had said 20 and Chris counted 27,000 -- presuming he took the time to count them all. That's not so unreasonable as it sounds.

By definition, a username is only allocated if an individual (person or company, don't matter) has access to the Server resources. So if this Server had 20 Web site accounts and 10 of them were businesses with, say... 2,500 people each who each had their own Email and/or FTP access... you see my point.

Or is my defintion of access incorrect?

Originally posted by Incognito
I'm sorry, but I agree with the others. The reason is because the question we hear all the time is "how many sites are hosted on the server"?

Why do people ask "how many websites are hosted on their server?" That's seems even less valuable than "how many usernames..." since one user can host multiples websites in subdomains and directories. I think you know what I'm trying to ask. How would you ask it so that I can double-check the answer, and make sure they don't put 800 people on the same server with me?