How many "typical" sites can you count on per server? Of course I know you have to play it by ear and watch performance, etc. but there has to be a number you can use to estimate.

Hmm?

sure.. 2
That's how many I have on one of my servers. They are fairly large and rely completely on php/mysql :D

But another server hosts smaller static HTML pages and can host hundreds of sites. It varies grately on your hardware and the type of sites it is hosting.

OK... 2. I guess I need about 150 servers then.

Yeah it all varies but when opening your servers to the world for general hosting there should be some sort of guess.

Is it 200? 400? Surely not 700 like ReadyHosting.

you can put 1000 sites on a server if you wanted to (and yes each would still load just fine)...

it all depends on how busy the sites are and what kind of resources they take up...

You should put them into basic plans.

For example you shouldn't put your Mega Super Duper Plans that give 20 MySql Databases on a server with another 100 Mega Super Duper Plans.

You need to basically watch your loads and average your growth by what your getting now. (Always expect the worst)

Hope this helps.

Thanks

Yes that helps. I have one pretty powerful server now and at the right time I will add another server. I will then balance the loads between the boxes as needed, etc.

I already have my server "breaking even" on costs but of course since I am a greedy bastard I want to move into profit-ville.

Yeah, you really can't even guestimate a number as you don't know what kind of clients you will be getting. All it takes is one fairly lage site to sign up, and you could find yourself only being able to fit a few more sites on it.

At least look at your stats. How much bandwidth do you get a month? How much useable storage do you have on the drives? Then take your hosting plans, and plan for how many would sign up with your plans and see how many you could fit. If you are offering a lot of high bandwidth/storage plans, you will obviously fit a lot less.

Same problem comes with load issues too. You could be hosting 300 kinda small static HTML sites no problem. Then one person signs up with a very busy message board using many of your system resources.. etc. It can be a juggling act :)

Yeah it is going to be quite a learning experince for me, that is for sure. Luckily I plan on having a seperate SQL Server machine which will take mucho load off my web server.

I just hope I don't screw it all up while learning.

We put 100-150 sites per server.

Thanks HP. I am assuming a normal distribution of clients for your numbers.

Originally posted by WonderMonkey
How many "typical" sites can you count on per server? Of course I know you have to play it by ear and watch performance, etc. but there has to be a number you can use to estimate.
How many people can you fit into a building?

How much water can you fit into a swimming pool?

Answer those 2 questions and you pretty much have your answer. ;)

Originally posted by hostingplex
We put 100-150 sites per server.
You must be busy managing your 40+ servers then. :D

Originally posted by hostingplex
We put 100-150 sites per server.
Server integity there depends on what you promise each of those accounts. If you've setup 150 sites where you promised 10GB disk space and 50 GB data transfer each, get ready for a wild ride. :D :eek:

Originally posted by Aussie Bob

You must be busy managing your 40+ servers then. :D

*shakes head* :D :D :D

Still toying with that one aren't you :laugh:

Originally posted by Aussie Bob

You must be busy managing your 40+ servers then. :D

You kill me. :D

I don't know what I expected to see first Bob. You explaining that the number of domains on a box is irrelevant, or you jumping on to Hostingplex again. You've got to stop watching Jerry Springer. ;)

Gary

Originally posted by Editor
I don't know what I expected to see first Bob. You explaining that the number of domains on a box is irrelevant, or you jumping on to Hostingplex again. You've got to stop watching Jerry Springer. ;)
I did both actually. :D:agree:

Originally posted by Aussie Bob

How many people can you fit into a building?

How much water can you fit into a swimming pool?

Answer those 2 questions and you pretty much have your answer. ;)
Bob,

I'll stick with your answer: 2 :D

Who needs crowded swimming pools or buildings?

-Matt

Acutally, we have a lot of servers. I haven't been in the datacenter for a while, so I couldn't tell you the exact numbers.

I know that we try to keep our servers running at optimal speeds, mix and match clients; both small sites and large. Then again, the servers aren't cheap and are built to handle extreme loads. IBM :)

Originally posted by hostingplex
Acutally, we have a lot of servers. I haven't been in the datacenter for a while, so I couldn't tell you the exact numbers.

I know that we try to keep our servers running at optimal speeds, mix and match clients; both small sites and large. Then again, the servers aren't cheap and are built to handle extreme loads. IBM :)
I aint touching that one. *looks at UmBillyCord* - "you wanna give it a shake"? :D

Thanks for the comments all. Here are my server stats. Also the sever is on its way and not up and running yet.

I have a Dell Poweredge 2650 with 1gb memory and a Xeon 2.2ghz processor. Three 36gb drives with Raid 5 which equates to 72gb space not counting OS, etc.

WonderMonkey --
That should do you good! 100 to 150 sites should easily be hosted on that server. Just always remember to monitor heavy users, etc.

Try to optimize the server as much as you can. Good luck!

Thanks HP. I am starting slow and only taking in about 10 sites to start out. Once I feel confident I will open it up to the world.

Originally posted by hostingplex
WonderMonkey --
That should do you good! 100 to 150 sites should easily be hosted on that server. Just always remember to monitor heavy users, etc.

Try to optimize the server as much as you can. Good luck!
Depends on what those 100 to 150 sites are doing. There's no magic number for sites/accounts/server. It depends on how much resources you promise [you make the choice on how you oversell etc] each account that goes onto the server.

Remember, you're like a virtual real estate agent that's subleasing an apartment complex. Each tenant you bring into the building pays you $X amount per month and for that payment, they get allocated a certian size apartment. The more they pay, the bigger their apartment is.

So when someone asks "how many tenants can we get into this building" and then someone replies with "you'll get 100 to 150 tenants" is not right. Depends on each of those tenants requirements.

Following on from my previous post -

Lets say you have your 20 floor apartment complex and you're signing up leases and handing out the keys. Donald Trump pulls up in his Limo and says he wants the top 15 floors. He opens a briefcase and hands over a wad of $$$$$ :pimp:

Now here's the $64.00 question -

"Does this 1 tenant have an impact on how many other tenants you can hold in your building"? :D

Maybe your servers should be populated by the size of the requirements that your tenants purchased and not by "100 to 150" tenants....?? ;)

I don't think Donald Trump needs webhosting though? or am I lost :>

Matt

Originally posted by Aussie Bob
Following on from my previous post -

Donald Trump pulls up in his Limo and says he wants the top 15 floors. He opens a briefcase and hands over a wad of $$$$$ :pimp:

[b]Now here's the $64.00 question -

"Does this 1 tenant have an impact on how many other tenants you can hold in your building"? :D

Oh Bob. If only it was that easy. :)

Originally posted by Aussie Bob
So when someone asks "how many tenants can we get into this building" and then someone replies with "you'll get 100 to 150 tenants" is not right. Depends on each of those tenants requirements.

Would you build an apartment block with no projections in mind? And then start knocking down walls and rearranging things, inconveniencing at least some folks?

And regardless of how much floor space you end up giving to someone, there is always going to be an implied limit to the number of occupants they can fill their space with.

We've seen what happens when the 6 P's aren't applied. :)

Gary

Originally posted by mattonline
I don't think Donald Trump needs webhosting though? or am I lost :>
Yes you are lost. :D

I was using an analogy of leasing an apartment complex and someone coming along [Donald Trump] and wanting to buy the top 15 floors. The apartment complex analogy works nicely when it comes to populating servers. Same business principles apply.

lol, nice analogy Bob :D

But that was exactally what I was saying earlier.. it all depends on who your clients are and what they require. You could easily fit 100 small sites on your server using your smallest plan, but what hapens when a big dog signs up and wants your biggest plan with most resources? your 100 sites may already be using most of those. So you either oversell and risk performance issues and such, or you buy a new server and juggle things around a bit, etc.

So no, there is no set numbers as it only takes one client to throw your whole plan out the window. The key is, be flexible and expect the worst :)

Originally posted by Editor
Oh Bob. If only it was that easy. :)



Would you build an apartment block with no projections in mind? And then start knocking down walls and rearranging things, inconveniencing at least some folks?

And regardless of how much floor space you end up giving to someone, there is always going to be an implied limit to the number of occupants they can fill their space with.

We've seen what happens when the 6 P's aren't applied. :)

Gary
It's an analogy Gary. Not meant to be taken literally like that. 1 tenant could take the whole building, or 250 tenants could take the building. Either way, the building will still operate, provided it is properly maintained. :)

I'd put two tenants in there and one bathroom, they can fight over it until one wins, survival of the fittest.

Matt

Originally posted by Aussie Bob

1 tenant could take the whole building, or 250 tenants could take the building. Either way, the building will still operate, provided it is properly maintained. :)

Okay, so carry on along the same lines of your anology mate. Would you attack either/or situation without some prior planning and projections. I know what you're saying in principle, but it only applies loosely at best. Perhaps a shopping centre complex would be a better anology, where it's not uncommon to drop a dividing wall to accomodate a varying size of retailer. But at the end of the day, if Kmart haven't put their money on the line for 1/4 of that complex up front, they are not going to be able to simply move in 6 months later.

Gary

Originally posted by Editor
Okay, so carry on along the same lines of your anology mate. Would you attack either/or situation without some prior planning and projections. I know what you're saying in principle, but it only applies loosely at best. Perhaps a shopping centre complex would be a better anology, where it's not uncommon to drop a dividing wall to accomodate a varying size of retailer. But at the end of the day, if Kmart haven't put their money on the line for 1/4 of that complex up front, they are not going to be able to simply move in 6 months later.

Gary
I think you're missing the point that you can't take certian aspects on an analogy and pick_them_and_pull_them into the discussion that the anology is representing. You can't then debate those aspects of the anaolgy in the same context as the discussion that the analogy represents.

It's just an anaology that demonstrates populating a server is dependant on the requirements of the tenants in that building/server. If you want to go with a "shopping center" picture, rather than an apartment complex picture, then that's ok. Yes, :D you can customise the requirements of each tenant better in a shopping center layout.

Oh and BTW, haven't you heard of folks buying several floors of an apartment building and then modifying the layout [yes, even knocking down walls ;)] to suit their needs. It's expensive, but some folks have too much money these days. :D

It's just urks me when someone says, "how many sites can a server hold", and someone answers, "you can get 100 to 150 sites per server". :eek: :rolleyes:

Here's one -

Host A populates a server with 100 accounts. Each account has been promised 5Gb disk space and 50GB data transfer.

Host B populates a server with 150 accounts. Each account has been promised 250MB disk space and 2GB data transfer.

Which server has the best chance of being more reliable?

Originally posted by Aussie Bob
It's just urks me when someone says, "how many sites can a server hold", and someone answers, "you can get 100 to 150 sites per server". :eek: :rolleyes:

Probably not as much as it irks me when someone says it is totally irrelevant. :)

But I didn't miss the point of your anology. I compared your entire analogy, and decided it was too far out. So I will stick with the shopping centre one. :-)

("center" Bob? You're going to have to drop the "Aussie" out of your handle soon. ;))

Gary

Originally posted by Aussie Bob
Which server has the best chance of being more reliable?

No, I'm not going there. Re-read the last thread that we went over that in. My answers will still be the same. :)

Gary

Originally posted by Editor
Probably not as much as it irks me when someone says it is totally irrelevant. :)
Where did I use those exact words?

Originally posted by Editor


No, I'm not going there. Re-read the last thread that we went over that in. My answers will still be the same. :)

Gary
Ok, what if they both have the same number of accounts, but promise different resources?? Wouldn't Host B's server have a "better chance" of being more stable??

Originally posted by Aussie Bob
Where did I use those exact words?

It was an anology Bob. :)

You have always maintained that the number of domains on a server is irrelevant (or words to the effect "not this argument again"). And that it's about the resources you promise relative to the money you can make on a box. And I have always maintained you're wrong. We'll get over it one day.

Cheers

Gary

Originally posted by Editor


It was an anology Bob. :)

You have always maintained that the number of domains on a server is irrelevant (or words to the effect "not this argument again"). And that it's about the resources you promise relative to the money you can make on a box. And I have always maintained you're wrong. We'll get over it one day.
What's to "get over"? It's a good discussion. :D

I agree. The main thing though, is we're both putting our methods to good use.

Gary

Aussie--

Which is why I mentioned to monitor use of heavy users. By this I meant exactly what you are trying to say :)

Originally posted by hostingplex
Aussie--

Which is why I mentioned to monitor use of heavy users. By this I meant exactly what you are trying to say :)
I thought you "ignored" me?? :confused: :D

I shoot for 90 - 100 sites per server, if we find one that's too resource-heavy I'll move it.


Oh yeah, HostOnce openly admit they put 1500 sites per server and use dual p4, 512 meg of ram and 40 gig hds. Now explain to me how they can offer Unlimited everything :D

Originally posted by UZIPP
I shoot for 90 - 100 sites per server, if we find one that's too resource-heavy I'll move it.


Oh yeah, HostOnce openly admit they put 1500 sites per server and use dual p4, 512 meg of ram and 40 gig hds. Now explain to me how they can offer Unlimited everything :D


Let me get this straight you put 90 sites at $5.00 month on a server and each site gets unlimted space and bandwidth:rolleyes:

Quote from your site
"our servers are on 100Mbps connections that can support as many sites as we can fit on our servers."

So for less than $500.00 a month Gross you can connect each server to a 100Mbps and stay in business? :confused:

I would love to see your business plan.

I hope you haven't made 1452 posts and advertised with "When your serious about reselling" because you know that's awful English right? Doesn't make you look serious at all.

It has worked just fine for me so far and considering I try to get very little business from here it does not really matter.

But I thank you for your review of my signature:D I will take it under advisement.;)

Originally posted by Naws
I hope you haven't made 1452 posts and advertised with "When your serious about reselling" because you know that's awful English right? Doesn't make you look serious at all.
Monte's only had that sig for a day or 2 now. Hardly comparable to offerring unlimited everything for 5 bucks a month. :D

OK I fixed it:D

Now you can go to bed happy. ;)

The "ultimate" way to handle this if you can would be to have multiple servers, one for each level of plan for the most part..

Server 1 - 100 basic plans
Server 2 - 50 standard plans
Server 3 - 10 Advanced plans.
Server 4 - You need a dedicated server fool!
Now.. client on server 1 starts pushing the limits, causing a load, You ask him to upgrade and move him to Server 2, and so on..

With 3 servers you have now provide a valid path of migration for each client.. By mixing clients across all 3 servers you reduce your ability to narrow down the problem children, and your ability to a single domains true load on a box. You also do not allow yourself the ability to provide more "room" for a single domain to spike.

We do something similar to this. Our top two plans come with features such as JSP which may be more resource intensive. For this reason they go on different servers than our lower plans.

Essentially our process has been to rank our plans though. We don't populate servers based on the number of domains, but rather the number of "units." For example, our low end plan may be 1 unit, where our 2nd plan may be 3. It is all based on past experience and statistics regarding what folks usually do when they sign up.

That being said, you always have to monitor your servers to account for sites that are out of your normal distribution. For business and capacity planning, you have to use averages... but for day to day operations, you must react to actuals.

The point being, you can't plan for actuals without knowing what they are going to be. Sure you may have a customer that uses all of the resources allocated, and then some. You won't know this until it happens though, and not knowing is not an excuse for not planning.

Thanks for the help all. I appreciate it. Good stuff.

cyberlot -> I plan on migrating towards something like that over time. Use my Win2k server(s) for those that need interactivity and a unix box for those that just want static sites.

be-hosted -> Interesting. I am going to remember that approach.

Originally posted by be-hosted
Essentially our process has been to rank our plans though. We don't populate servers based on the number of domains, but rather the number of "units." For example, our low end plan may be 1 unit, where our 2nd plan may be 3. It is all based on past experience and statistics regarding what folks usually do when they sign up.
Well said. Anyone who populates a server simply on the "number of domains", need their head read, IMO. :D

I agree Bob and all I was wanting where some "In General" type numbers. Mixed in this thread where those types of responses.

Originally posted by Aussie Bob

Well said. Anyone who populates a server simply on the "number of domains", need their head read, IMO. :D

Nice try Bob. :) But firstly, be-hosted started offering reseller plans 3 weeks ago so I am assuming his projections still relate back to the number of domains on a server anyway. But if not, their reseller plans have a set limit of domains each. Correct me if I'm wrong...but I think there's some consideration being given to the "number of domains" on a server there as well.

Cheers

Gary

The average site requires less then 5 mb of space, many sites are larger however, so its hard to say. I think 50 - 100 sites is a good bet on an average server...

Are there any hosts here that populate a server as per the "number of domains" and not what they promise each account??

Or, more to the point...does anyone else disregard the number of domains on a server? ;)

Gary

Originally posted by Editor
Or, more to the point...does anyone else disregard the number of domains on a server? ;)

Gary
Typical Queenslanders. Don't know when to quit. ;)

So what's the magic number to domains per server?

Originally posted by Aussie Bob
Typical Queenslanders. Don't know when to quit. ;)

I'm a Sandgroper on holiday. ;)

It's not a matter of quitting though Bob. I'm sharing my opinion based on a few years experience of what works for us. But on top of that, I'd be very surprised if many would openly disagree that you need to take the number of domains on a server in to consideration at all times.

There's no magical number. It depends largely on the server specs. But IMO it has to be one of the *first* considerations. You take the cost and specs, and project a number of domains that the server could handle (let's say 250). You devise a set of plans that takes in to consideration room for stats/logs/mail/etc, and which will return a profit that then goes on to cover other operating overheads.

Sure, a site might come along that has some intensive requirements, and it pays its way accordingly. So the number of domains that end up going on to that box could drop significantly. Or you might hit that 250 and decide it could handle more.

But without considering upfront just how many domains you are going to put on a server you are running blind. It is the only constant to work on, *if* each plan stays within the projected disk space and data transfer. Anything *after* that you work on as it comes.

We'll have an A4 page of notes/projections in place before purchasing/populating a new server. If we have a server of similar specs in the same datacentre we're considering, then we already have plenty of track records to work on.

That's my take on it. :)

Gary

Originally posted by Aussie Bob
Are there any hosts here that populate a server as per the "number of domains" and not what they promise each account??

Not sure why you can't understand many places limit servers in terms of domains and not resource use. As I stated before we never go over 300 on dual Xeon, 1 GB RAM servers. That is the limit. We also have been in business long enough to know our customer base and who we attract. We do not get script kiddies and cheap skates wanting to cram 40 MySQL databases running JSP on a $10/mo account. So our chances of getting high resource guys is small. Once we get them, we move them.

We set a target amount of revenue per server. 300 sites/server has averaged out just what we need. Sure, 90% of our servers could have 500 sites on them. But why? We meet our target and our customers have better uptime because there is a smaller propensity of someone writing bad scripts or smaller change of getting an abusive customer. Also, it is good advertising to post, in writing, what you promise your customers the max you would allow. Once someone learns crowed servers are bad, they will look for this. Trust me.

Before you ever look at how many sites per server either by # or resource as the deciding factor, you need to plan how much money you need to make per server. Simply for worse -case senerio planning. If you don't plan right, you could be losing money and would need to re-evaluate your business madel or plans pricing.

Originally posted by UmBillyCord
Once someone learns crowed servers are bad, they will look for this. Trust me.

We're getting back now to most prospective clients asking "how many domains per server", like it was before this years' blowout in the unlimited domains idea. Previously, folks would tear apart a hosting model that supported it because of the heavy restrictions that need to be in place. Donhost have been doing it successfully for a few years now, but not everyone can work within their model.

Gary

Originally posted by UmBillyCord
We set a target amount of revenue per server. 300 sites/server has averaged out just what we need.
(1). Ok. What's the average disk space and data transfer do you promise each of those 300 accounts?

(2). In your opinion, is there a real difference if you promised each of those accounts 1GB disk space and 25GB data transfer or if you promised each of those accounts 100MB disk space and 2 GB Data transfer?

(1) I thought he had Bob. That was that long thread where we ended up posting server loads/etc wasn't it? I believe UBC answered with 300 there too?

(2) There is a "real" difference if you are comparing one domain to "unlimited" domains within that space (in either scenario). DNS calls. Hands on management time (fault finding, etc). mySQL connectivity. Email. FTP.

Cheers

Gary

Originally posted by Editor
(1) I thought he had Bob. That was that long thread where we ended up posting server loads/etc wasn't it?
[edit] Post was replied to.
I believe UBC answered with 300 there too?
Yep. I just wanted to know on average what each of those 300 accounts were promised as far as disk space and data transfer usages.

Originally posted by Aussie Bob
Maybe you could find his answer from the thread if you believe UBC posted it. :)

Link to the thread was the question presented in?

Gary

Originally posted by Editor
Link to the thread was the question presented in?
Crikey Gary, :D that was a fair while ago. I'm too busy right now to hunt for it. I do know that I didn't get an answer. That's ok. I thought it might have been too sensitive a question to ask. But UBC is completely anonymouse here anyways, so I saw no harm in asking. :)

Originally posted by Aussie Bob
I'm too busy right now to hunt for it. I do know that I didn't get an answer.

You raised the point Bob...so you should have at least checked. :)

Like I said...I seem to recall UBC did answer and that it was quite some time after you presented the question. But, if you can find the thread that'd be helpfull.

Cheers

Gary

Originally posted by Editor
You raised the point Bob...so you should have at least checked. :)
I'm searching now. I know for sure that he didn't answer that one.
Like I said...I seem to recall UBC did answer and that it was quite some time after you presented the question. But, if you can find the thread that'd be helpfull.
I'm looking, but the current question still stands. Is there a rule about only answering it the one time? ;)

Originally posted by Aussie Bob
Is there a rule about only answering it the one time? ;)

Of course not...but you're saying it was asked and ignored already. ;)

Gary

Originally posted by Editor
Of course not...but you're saying it was asked and ignored already. ;)
Yep, it was. Hunting through all UBC's posts now.....:eek:

Ok :D I found his answer (http://webhostingtalk.com/showthread.php?s=&threadid=74360&perpage=15&highlight=300&pagenumber=6) [apologies - my question to UBC was replied to. :blush: ]

UBC said -
To let you know, no matter how low the server use is, we always stop adding at 300 sites No matter what.
And I then asked -
Ok - 300 accounts per server. What is the average promised disk space and data transfer/account on that server??
UBC then replied with -
Servers run 80 GB HDs (older ones - 40s). Our basic plans start from 100 MBs up to 500 MBs.....Bandwidth varies and is not relevent.
So that's 250MB/account on average?

Then UBC said -
We do not determine the # of accounts based off of resources. We base it on target revenue/server and hard numbers derived through time.
Same here. US$1,000/mth/server for us. You'd be making much more, I imagine. But you would have calculated backwards from the resources to make sure the server can hold those clients who were promised certian resources?? Setting up your plan prices to attain a certian level of revenue/server would also be considered.

You're starting from predetermined points to get your "300 domains per server". So you have limited your domains as to what they can use as far as disk space and data transfer. That's how you populate your servers and that's fine. :)

So I don't see how you can say -
[b]Not sure why you can't understand many places limit servers in terms of domains and not resource use.
When it's clear that you predetermine the what each of those 300 accounts can use as far as disk space and data transfer when you create the plans in the first place. You couldn't have each of those 300 accounts promised 5GB disk space and 25Gb data transfer??

Originally posted by Aussie Bob
Same here. US$1,000/mth/server for us.

Doesn't work Bob. Unless I'm missing something.

300 x Plan 1, or
50 x Plan 4
etc

They should always work out closely in revenue, and be a viable situation. Simple as that. If you put 10 of your Plan04 on a box (which arrives at the same disk space/data transfer at what you mention above), you take a 50% hit and still with the same overheads. Would you then put 20, in order to meet your projections? All of which have unlimited domains at their disposal (which was actually still my point...sorry for the deviation).

Gary

When it's clear that you predetermine the what each of those 300 accounts can use as far as disk space and data transfer when you create the plans in the first place. You couldn't have each of those 300 accounts promised 5GB disk space and 25Gb data transfer??

Bob. You split hairs. And you combined a few different methods to work your arguement. No one said we do not use resources to assist in determining what the ideal number is for us to limit our servers. What I am saying and I think Gary is saying, is that resource use once the server is running, is an after thought. Seldom do we get a server that will tap out on resource use before we hit 300 sites. So you see, the 300 site limit is what we use. It is even coded into our software. If a site causes high resource use which will effect that limit, we move them to a server that shares resources with others that are high resource. Once that is done, the server is back on track to meet our domain limit. Also, we have not raised or changed prices since we started. So it is not hard to figure out our typical ratio of plans and how easy it is to meet our revenue goal.

As stated before, our reasoning is simple:
1) Marketing. Placing a limit will provide additional reassurances to our users we will not be someone like Hostonce or some other place that crams doamins. People willlearn over time, if you cram servers, you have a greater chance to have something go wrong.
2) We have no need to place more accounts on our servers. We meet our goals and that is all that matters.

Bob, you seem to miss the one key point. You are assuming everything will be smooth sailing when using resource use as a determining factor. Fine. But what happens when that fails. Your users suffer. If you live in a town of 500 people and a town of 5000, which one will have a better change to have a murder? Well, if you have a server with 500 sites vs. 5000, then of cource the chance of someone writing bad scriopts, etc... is greater. Bottom line.

Also, I don't think anyone is saying one method is better then another. We just choose to use a static determining factor then a dynamic. Resource use is dynamic.

Originally posted by Editor
Doesn't work Bob. Unless I'm missing something.

300 x Plan 1, or
50 x Plan 4
etc

They should always work out closely in revenue, and be a viable situation. Simple as that. If you put 10 of your Plan04 on a box (which arrives at the same disk space/data transfer at what you mention above), you take a 50% hit and still with the same overheads. Would you then put 20, in order to meet your projections? All of which have unlimited domains at their disposal (which was actually still my point...sorry for the deviation).
Most accounts on our servers are the Plan 1 @ $25/mth for 1Gb disk space and 10GB data transfer. That's 40 Plan 1s per server, which equals US$1k/mth in revenue.

Not sure where you get "300 x Plan 1". :confused:

That was an eg (example), trying to work out where you were going with only making these projections based on revenue. Looking at it your way, with your plans, doesn't it mean you would only put 10 Plan 04's on a server and therefore raise half of that 1K per server? Which is alleviated by starting with a constant (number of domains) and working from there.

Cheers

Gary

Originally posted by UmBillyCord
Bob. You split hairs.
Coming from you UBC, that's a compliment. :)
And you combined a few different methods to work your arguement. No one said we do not use resources to assist in determining what the ideal number is for us to limit our servers. What I am saying and I think Gary is saying, is that resource use once the server is running, is an after thought. Seldom do we get a server that will tap out on resource use before we hit 300 sites. So you see, the 300 site limit is what we use. It is even coded into our software. If a site causes high resource use which will effect that limit, we move them to a server that shares resources with others that are high resource. Once that is done, the server is back on track to meet our domain limit. Also, we have not raised or changed prices since we started. So it is not hard to figure out our typical ratio of plans and how easy it is to meet our revenue goal.

As stated before, our reasoning is simple:
1) Marketing. Placing a limit will provide additional reassurances to our users we will not be someone like Hostonce or some other place that crams doamins. People willlearn over time, if you cram servers, you have a greater chance to have something go wrong.
2) We have no need to place more accounts on our servers. We meet our goals and that is all that matters.

Also, I don't think anyone is saying one method is better then another. We just choose to use a static determining factor then a dynamic. Resource use is dynamic.
We have similar methods to populate a server. You place 300 accounts per server and guarantee them "X" amount of disk space and "X" amount of data transfer/mth. You predetermine the prices of those plans based on their usage and this backs into your revenue per server and cashflow projections to run your business. You need a certian level of revenue per server to make your business model to work. You have also predetermined that those 300 accounts using their resources, won't kill the server.

From memory, I think you said you cheapest plan was $10/mth. That would be minimum revenue per server of 3k/mth. I think you're coloing your own servers that you probably build for $500/server. Maybe you pay $700/mth/rack. I'd guess that you're using 2U servers.

My point here is - you use predetemined usage limits per account per server to arrive at your plan sizes and prices. It maybe 300 domains per server, but you have predetermined what each of those accounts can use and how much revenue you will derive from each account. That's a fair way from populating the server on the basis of the number of domains only.
Bob, you seem to miss the one key point. You are assuming everything will be smooth sailing when using resource use as a determining factor. Fine. But what happens when that fails. Your users suffer. If you live in a town of 500 people and a town of 5000, which one will have a better change to have a murder? Well, if you have a server with 500 sites vs. 5000, then of cource the chance of someone writing bad scriopts, etc... is greater. Bottom line.
Ok :D now we're talking about "chances" here -

Server A has 250 accounts. Each account has been promised 5GB disk space and 25GB data transfer.

Server B has 250 accounts. Each account has been promised 100MB disk space and 2GB data transfer.

Which server has the best chance [there's that word again ;)] of superior performance?

I'd go out on a limb here :) and say that Server B has the best "chance" of delivering superior performance. Would this be a safe assumption?

Originally posted by Editor
That was an eg (example), trying to work out where you were going with only making these projections based on revenue. Looking at it your way, with your plans, doesn't it mean you would only put 10 Plan 04's on a server and therefore raise half of that 1K per server? Which is alleviated by starting with a constant (number of domains) and working from there.
Projections based on revenue and based on what type of plans will go on the server. Most will be the $25/mth plans. We have a lot of room to move if we have a couple of larger accounts on the server. 40 Plan 1 accounts at 100% disk space utilisation is only 40GB. They're 80GB drives. So at 100% utilisation, there's only 50% total utislisation of the server's hard drive. If we have larger accounts [Plan 4 to 5], then they take up the rest of the space, if needed.

Remember, we're talking about 100% client utilisation of their disk space. You would have a figure on what the true disk space utilisation is based on your experiences. It's a long way from 100%. :)

UBC, 2 more questions -

(1). If you had to reduce the cost of your plans, would this increase the number of accounts per server?

(2). If you had to increase the resources promised to each account, would this decrease the number of accounts per server?

Have you been drinking Bob?

Originally posted by Aussie Bob
Server A has 250 accounts. Each account has been promised 5GB disk space and 25GB data transfer.

Server B has 250 accounts. Each account has been promised 100MB disk space and 2GB data transfer.

Which server has the best chance [there's that word again ;)] of superior performance?[/B]

How do you arrive at this comparison? 250x5GB and 250x25GB with Server A? More like:

Server A has 5 accounts. Each account has been promised 5GB disk space and 25GB data transfer.

Server B has 250 accounts. Each account has been promised 100MB disk space and 2GB data transfer.

In which case Server A would most likely offer better performance given that those 5 domains could be managed much more efficiently than 250.

Which just happens to relate directly to the number of domains on a server. :)

Gary

coloing your own servers that you probably build for $500/server.

Man. I am insulted. It is amazing that people think RS, low-spec craps the norm. We run dual Xeon, RAID, min 1 GB of RAM 1Us with hot swap drives. Show me where we can get those for less then $2000.

You have also predetermined that those 300 accounts using their resources, won't kill the server.

Hell, Bob. 500 - 800 wouldn't kill our servers. At peak times, 95% of our shared servers will never run over .20 at a 5 minute avergae.

I'd go out on a limb here and say that Server B has the best "chance" of delivering superior performance. Would this be a safe assumption?

Of course. Traffic, not space will screw with resources. If you think about it, space is part of traffic.

Gary, we're using the "town analogy" that UBC said -
If you live in a town of 500 people and a town of 5000, which one will have a better change to have a murder? Well, if you have a server with 500 sites vs. 5000, then of cource the chance of someone writing bad scriopts, etc... is greater. Bottom line.
Hence in UBC's context, I asked -
Server A has 250 accounts. Each account has been promised 5GB disk space and 25GB data transfer.

Server B has 250 accounts. Each account has been promised 100MB disk space and 2GB data transfer.

Which server has the best chance [there's that word again ] of superior performance?
Can you answer the question in the context of UBC's "town" analogy?

My answer is Server B has the best "chance" of delivering better performance. Remember Gary, UBC introduced the "town" analogy, not me. :)

Originally posted by Aussie Bob
UBC, 2 more questions -

(1). If you had to reduce the cost of your plans, would this increase the number of accounts per server?

(2). If you had to increase the resources promised to each account, would this decrease the number of accounts per server?

1) Why would I need to do that? That is bad planning if we do. If we had to release a small commodity based plan for some reason, we would do it as a new company. Then we would lower our costs in server purchases and other factors.

2) It is called long term planning. We have a minuim shelf life of two years for every server we bring online. This means we will not need a larger HD or new motherbaord, etc... for two years. So basically we plan what resource requirements will be like in the future. We are prepared to double all resources with our next web site roll out. Doubling resources doesn't mean we will all of a sudden allow unlimited DBs, or JSP on low end plans for example. Doubling space, BW, e-mails are really just window dressing anyway.

How so Bob? UBC says "If you live in a town of 500 people and a town of 5000", and you liken that "If you're on a server with 5000 domains and a server with 5000 domains". I've missed your point, and asked if you've been drinking. :)

Gary

Originally posted by Editor
How so Bob? UBC says "If you live in a town of 500 people and a town of 5000", and you liken that "If you're on a server with 5000 domains and a server with 5000 domains". I've missed your point, and asked if you've been drinking. :)

Gary

I missed that one too. :)

Hence in UBC's context, I asked -

No, you then added to it. You then asked if that town with 500 was in Compton (LA, CA) or in small town Minnisota. Im saying keep the test group the same. We are talking # of domains. So don't throw resource variables into the mix.

;)

UBC's analogy claimed that the bigger the town, the more chance you've got of being murdered. Fair enough. No arguments there.

My following question asks which server has the best CHANCE of delivering the best performance. You missed that UBC's analogy was based around the word "chance". You use that word a LOT Gary, when discussing this matter.

This is the context. My question below uses that same context ["chance"] to determine which server will be the most reliable.

Yet noone will/can answer -
Server A has 250 accounts. Each account has been promised 5GB disk space and 25GB data transfer.

Server B has 250 accounts. Each account has been promised 100MB disk space and 2GB data transfer.

Which server has the best chance [there's that word again ;)] of superior performance?
I see no reasons whatsoever that the question cannot be answered. It's a simple question based around the context of "chances" as introduced by UBC in his town analogy.

My answer is Server B. Yet both servers have the exact same domains. How can 1 server be more reliable than the other and have the exact number of domains?

You know Bob...it really is pointless us going over this every time someone asks for advice about how many domains to plan on a server. You're always going to say it doesn't matter, and I'll say it does. You don't take it into consideration and you have no control over how many, and I do. No comparison or analogy is going to work.

We should agree to disagree just once in these types of threads, and be done with it. :)

Gary

Originally posted by UmBillyCord
No, you then added to it. You then asked if that town with 500 was in Compton (LA, CA) or in small town Minnisota. Im saying keep the test group the same. We are talking # of domains. So don't throw resource variables into the mix.
heh. God forbid that anyone throw resource variables into the mix. *cough* Gary - shopping centers / Apartment buildings....:)

It doesn't look like anyone is going to answer, Server A or Server B. :eek:

Originally posted by Aussie Bob
I see no reasons whatsoever that the question cannot be answered.

Read it again Bob. You're comparing apples to oranges, to the extreme.

My answer is Server B. Yet both servers have the exact same domains. How can 1 server be more reliable than the other and have the exact number of domains?

In your analogy, I dont know...especially when Server A is 250x25GB/250x5GB, and the other one is somewhat more realistic. :)

Gary

Originally posted by Editor
You know Bob...it really is pointless us going over this every time someone asks for advice about how many domains to plan on a server. You're always going to say it doesn't matter, and I'll say it does. You don't take it into consideration and you have no control over how many, and I do. No comparison or analogy is going to work.

We should agree to disagree just once in these types of threads, and be done with it. :)
Both servers have the exact same number of domains. I am simply asking which server has the best "chance" of being more reliable. That's all. Forget previous discussions - just keep the focus on this issue at hand. I'll push hard for an answer here Gary. I'm not being arrogant or deceptive. I'd just like an answer to the question. If you don't want to answer it, just say so. :)
Server A has 250 accounts. Each account has been promised 5GB disk space and 25GB data transfer.

Server B has 250 accounts. Each account has been promised 100MB disk space and 2GB data transfer.

Which server has the best chance [there's that word again ;)] of superior performance?

Originally posted by UmBillyCord
1) Why would I need to do that? That is bad planning if we do. If we had to release a small commodity based plan for some reason, we would do it as a new company. Then we would lower our costs in server purchases and other factors.

2) It is called long term planning. We have a minuim shelf life of two years for every server we bring online. This means we will not need a larger HD or new motherbaord, etc... for two years. So basically we plan what resource requirements will be like in the future. We are prepared to double all resources with our next web site roll out. Doubling resources doesn't mean we will all of a sudden allow unlimited DBs, or JSP on low end plans for example. Doubling space, BW, e-mails are really just window dressing anyway.
Good answers, but they didn't directly answer my question. I'm not interested in why you would or wouldn't need to do that, or your planning. I'm just interested in the exact answers for -
(1). If you had to reduce the cost of your plans, would this increase the number of accounts per server?

(2). If you had to increase the resources promised to each account, would this decrease the number of accounts per server?
They're "Yes" or "No" type answers too. :)

Originally posted by Aussie Bob
I'll push hard for an answer here Gary.

Then present a half reasonable question Bob. Server B is fairly standard setup. Server A has you away with the fairies. ;)

What do you expect me/anyone to base an answer to a ridiculous question like that on?

Gary

Originally posted by Editor
Then present a half reasonable question Bob. Server B is fairly standard setup. Server A has you away with the fairies. ;)

What do you expect me/anyone to base an answer to a ridiculous question like that on?
Far from a rediculous question mate. Both servers have the exact number of domains, yet Server B is the most reliable server. Server A was setup by a host overselling like crazy with cheapo plans. Server B was populated by a host who knows when a server has reached its limits. Both servers are real world scenarios. Not "away with the fairies" like you mention. :)

Hence this supports my previous statement (http://webhostingtalk.com/showthread.php?s=&threadid=74360&perpage=15&highlight=300&pagenumber=6) made on 09-19-2002 01:09 PM [which you always misquote ;)] that -
My statement is - "The number of domains on a server is not always relevant to that server's actual performance."

Bob, I'm now absolutely convinced you're drunk. At no stage has overselling been a part of this discussion...except now as justification for your line of questioning. If you had dropped that in earlier, I would set you straight two pages ago. :)

Can we keep this on topic now?

Gary

Originally posted by Editor
Bob, I'm now absolutely convinced you're drunk. At no stage has overselling been a part of this discussion...except now as justification for your line of questioning. If you had dropped that in earlier, I would set you straight two pages ago. :)

Can we keep this on topic now?
Please keep off the "drunk" comments too. They're not appreciated and greatly diminish your arguments.

Gary, it was plainly obvious that Server A was oversold to the hilt. That's my point. Thank you for pointing that out. The host was selling cheapo plans and promising way too much in resources.

The host has 250 domains on the server BUT the each of those domains is promised way too much disk space and data transfer [not that this happens in our wonderful industry ;)]

Therefore that server would have a higher "chance" of lower performance. Hence my statement -
"The number of domains on a server is not always relevant to that server's actual performance."
Is the above statement right or wrong?

Mods - this discussion might get a bit heated, but that's ok. Friction is a good and positive thing. :)

How have you assumed that the "5GB disk space and 25GB data transfer" plans are "cheapo". If you had added $5/mth to both, then you would have had a point to your questions, but still would have failed to have anything to do with the topic at hand.

Sorry about the "drunk" comment mate. It's a saying, as I would have assumed you'd know. "What, are you drunk or something?" expresses an immense amount of disbelief. :)

So no, in my opinion your statement is still wrong. Definitely wrong in each of the scenarios/analogies/comparisons you have presented. The number of domains would be the first consideration when projecting target performance levels, and the plans structured around that accordingly.

Cheers

Gary

Originally posted by Editor
How have you assumed that the "5GB disk space and 25GB data transfer" plans are "cheapo". If you had added $5/mth to both, then you would have had a point to your questions, but still would have failed to have anything to do with the topic at hand.
My statement still stands. -
"The number of domains on a server is not always relevant to that server's actual performance."
Sorry about the "drunk" comment mate. It's a saying, as I would have assumed you'd know. "What, are you drunk or something?" expresses an immense amount of disbelief. :)
No worries. :)
So no, in my opinion your statement is still wrong. Definitely wrong in each of the scenarios/analogies/comparisons you have presented. The number of domains would be the first consideration when projecting target performance levels, and the plans structured around that accordingly.
IMO, that's not accurate. Server B is the better populated server, even though the same number of domains are on each server. Consideration is given to the actual promised USAGE of each account on the server and not just the actual number of domains. Therefore - "The number of domains on a server is not always relevant to that server's actual performance."

That server has a better "chance" [your word] of better performance than Server A which the host has oversold to the hilt. That's a no brainer.

Exact same spec servers - same rack in the datacenter - A server can have 1 domain that pushes 1000GB/mth [complex CGI heavy search engine with 15 minute load average of 8.5] and another server can have 100 domains that are barely push loads past .1 over 15 minute average. Therefore - "The number of domains on a server is not always relevant to that server's actual performance."

To prove my statement wrong, you would need to present conclusive evidence that clearly states - "The number of domains on a server is always relevant to that server's actual performance."

Please note the important differences in "not always" and "always". Sorry if this is "splitting hairs", but I feel it's paramount to this discussion. Can you produce this concrete evidence that conclusively proves - "The number of domains on a server is always relevant to that server's actual performance."?

I'd like to keep this discussion focussed entirely on that aspect and not wonder all over the place [as per your request]. :)

Bob is this the same way you explain to people around the BBQ what you do for a living?

They prob think your in real estate LOL.

Keep up the good work Bob. :)

Dont oversell would be my answer and split it compared to the plans you want to offer. Dont put the big plans all on one server.

if you Cpu or Memory gets low and the server becomes un reliable upgrade or move some clients to another server. Extra ram should be sweet.

With the power edge that you have even though most wont tell you how many domains they really host on a server it would be close to 200-300 depending alot of course on the usage.

If you have heaps of small sites and dont use to much you can push the 400-500 easy.

Kind regards
Phillip

Originally posted by Aussie Bob
Consideration is given to the actual promised USAGE of each account on the server and not just the actual number of domains.

Not in your example it hasn't. Unless you can demonstrate that. Do you know of any servers with 250 accounts and each account has been promised 5GB disk space and 25GB data transfer?

So you're still stuck on apples and oranges mate. :)

Gary

Originally posted by Editor


Not in your example it hasn't. Unless you can demonstrate that. Do you know of any servers with 250 accounts and each account has been promised 5GB disk space and 25GB data transfer?

So you're still stuck on apples and oranges mate. :)

Gary
lol. Ok, no worries then. :D

My statement stands unopposed -
"The number of domains on a server is not always relevant to that server's actual performance."
If you or anyone [UBC ??] wish to prove that statement incorrect, you'd need to provide conclusive proof that -

"The number of domains on a server is always relevant to that server's actual performance."

And that's something I'd like to see. ;)

Originally posted by Aussie Bob

Good answers, but they didn't directly answer my question. I'm not interested in why you would or wouldn't need to do that, or your planning. I'm just interested in the exact answers for -

They're "Yes" or "No" type answers too. :)

Bob. Not everything is black and white. Ex..

If you murder someone, would you use a knife or gun. Well, the answer would be you would never murder someone so why worry about it?

Also, you keep saying no one answered your Server A vs. Server B. Please read my post again. I clearly did answer it.

Is the above statement right or wrong?

Of course not. You covered your ass by using a qualifier "not always". Listen to what we are saying:

The number of domains on a server is not always relevant to that server's actual performance. HOWEVER, the increase in the number of domains on a server will also increase the chances of someone affecting the server's actual performance. There is a positive corolation between the two. If you do not see this, then you may want to check the water you drink. :)

Originally posted by Choppy
Bob is this the same way you explain to people around the BBQ what you do for a living?
I don't even bother these days. If they're not online and don't have a vague understanding of the internet, I don't even bother to explain how hosting works. I used to, but now I just give up. :D
They prob think your in real estate LOL.
Well, we're really just Virtual Real Estate agents. ;)
Keep up the good work Bob. :)
I love this argument of domains per server. I'm still waiting for the below statement to be proven wrong -
"The number of domains on a server is not always relevant to that server's actual performance."

Originally posted by UmBillyCord
Of course not. You covered your ass by using a qualifier "not always".
UBC. That was my statement back several months ago. It still stands. The "covered your ass" is what we call being accurate when making statements. Hardly covering anything up mate. :)
Listen to what we are saying:

The number of domains on a server is not always relevant to that server's actual performance.
Thank you. :)
HOWEVER, the increase in the number of domains on a server will also increase the chances of someone affecting the server's actual performance. There is a positive corolation between the two. If you do not see this, then you may want to check the water you drink. [/B]
The water here is fine. UBC, I see that and agree with that. The "chances" [there's that word again ;)] are higher.

The problem is, Gary has been misquoting my recent statement that I made ages ago about this issue, saying that I claimed that the number of domains on a server has no relation to the server's performance. :eek:
Posted by Gary-
Probably not as much as it irks me when someone says it is totally irrelevant.
Posted by Gary -
You have always maintained that the number of domains on a server is irrelevant....
However my statement was -
"The number of domains on a server is not always relevant to that server's actual performance."
Splitting hairs?? Hardly. Covering my ass by using the words "not always" ?? Nope. It's just being precise. Thank you. :)

Splitting hairs?? Hardly. Covering my ass by using the words "not always" ?? Nope. It's just being precise. Thank you.

Precise? Precise is using the word "never". "Not always" leaves room to play both sides of the fence. That is what I mean by covering your ass. Wasn't meant to offend.

Originally posted by UmBillyCord
Precise? Precise is using the word "never". "Not always" leaves room to play both sides of the fence. That is what I mean by covering your ass.
I would disagree with that. A precise statement can be made using the term "not always". It means that it's not black and white, right or wrong. It's as precise as I can make it. I think you did say somewhere in this thread, that it's not a black and white issue. Many variables to be considered. :)
Posted by UBC -
Bob. Not everything is black and white. Ex..
Wasn't meant to offend.
NO worries about offending me. I love these types of discussions. I learn a lot. :) Friction is good.

Ok. :D So there would be general agreement with -

"The number of domains on a server is not always relevant to that server's actual performance. HOWEVER, the increase in the number of domains on a server will also increase the chances of someone affecting the server's actual performance."

Gary, is that a Yes or a No?

Well, we've come along way today then Bob. :) Not so long ago you didn't understand how the number of domains affects a server at all. One step at a time...so yes I'll come at that in a very general sense (especially the first part).

Cheers

Gary

Originally posted by Editor
Well, we've come along way today then Bob. :) Not so long ago you didn't understand how the number of domains affects a server at all. One step at a time...so yes I'll come at that in a very general sense (especially the first part).
Always understood that mate. A server with 5000 domains is got to have problems. There's extremes to the argument. Like UBC said, it's not black and white. Hence I've maintained that -
"The number of domains on a server is not always relevant to that server's actual performance.
I do like -
"The number of domains on a server is not always relevant to that server's actual performance. HOWEVER, the increase in the number of domains on a server will also increase the chances of someone affecting the server's actual performance."
Now we just have to nail down the % "chance".....;) :laugh:

After reading all 8 pages of this thread, I have come to a conclusion!

Aussie Bob
Editor
UBC

Have all added drastically to their number of post if the past few days.

CyberSorcerer

I need a new space bar... :)

Gary

Originally posted by Cyber-Dock
After reading all 8 pages of this thread, I have come to a conclusion!

Aussie Bob
Editor
UBC

Have all added drastically to their number of post if the past few days.

CyberSorcerer
Actually :D with civilised discussion, we arrived at -
"The number of domains on a server is not always relevant to that server's actual performance. HOWEVER, the increase in the number of domains on a server will also increase the chances of someone affecting the server's actual performance."
We just have to figure out the % chance now. That's the next discussion. ;)

Originally posted by WonderMonkey
Thanks HP. I am starting slow and only taking in about 10 sites to start out. Once I feel confident I will open it up to the world.

How do you figure to get only 10 sites? I mean, If youll publish your hosting site on internet, you wont be able to control the amount of costumers. After first 10 you could tell them you arent accepting any more, but that would be bad for company reputation, since people would start asking themselves why arent you accepting additional costumers, wouldnt it?

So, How do you think of getting 10 costumers?

rackyou -> The initial 10 sites come from a combo of my personal ones and a few I already manage for others. I am a developer by trade and have done a few sites with a decent database back end. I will put those and a some other statics on there for a while to make sure I am doing things correctly. When I am sure, I will then open it up.

Originally posted by Cyber-Dock
After reading all 8 pages of this thread, I have come to a conclusion!

Aussie Bob
Editor
UBC

Have all added drastically to their number of post if the past few days.

CyberSorcerer

Yes. And with this post here, you have added to yours. ;)

Hm.. How much are you paying for the server? I guess you have a dedicated server leased somewhere, right?
So, do these sites you have from your clients at least cover your expenses?

Oh hello - this has got to be one of the silliest threads for a while. But thanks Aussie Bob for giving us your costings - makes for interesting reading. Basically Aussie Bob's arguement is flawed for this simple reason. He offers unlimited domain hosting accounts using whm/cpanel. So where he says each server holds about 40 of his $25 per month reseller accounts, this is not the number of accounts that the server will end up holding. Each of these account holders is a reseller with thier own WHM from which they can create *unlimimited* domain name based cpanel accounts. So each server will not have only 40 cpanel accounts on it but an unspecified, unknown number of cpanel accounts.

As someone who also offers multi-domain hosting accounts, I like to have a target in my mind of how many cpanel accounts I think is reasonable to have on a server - which should roughly correlate to how many virtual hosts apache will be running.

Originally posted by freakysid
Oh hello - this has got to be one of the silliest threads for a while.
I simply couldn't be bothered to debate this one again, nor do I give a damn. :rolleyes:

What's been said, has been said and still stands -
"The number of domains on a server is not always relevant to that server's actual performance. HOWEVER, the increase in the number of domains on a server will also increase the chances of someone affecting the server's actual performance."
BTW, watch those "unlimited subdomains" and "unlimited POP boxes" too. Wouldn't want to get 100,000 subdomains or 100,000 POP email accounts on a server now would we?? Well, you do promise "unlimited" resources here. ;) :emlaugh:

For someone who doesn't give a damn you have a lot to say! I am just pointing out that your Plan A and Plan B arguement doesn't at all gel with your business model. Your arguement is that plan B is the superior way to populate a server, yet because you offer multiple domain hosting, in reality your business model resembles something more like plan A. And yes, as you have pointed out, there are always risks of offering unlimited anything - be it unlimited domains, subdomains or mailboxes. All i can say is that in reality I massively undersell my servers (that is massively underutilize their capacity) - and you probably do too, and many good web hosts probably do.

I'm just pointing out that this 40 accounts per server thing of yours does not equate to 40 cpanel accounts. That's not true because each customer can create many accounts under their WHM.

Originally posted by freakysid
I'm just pointing out that this 40 accounts per server thing of yours does not equate to 40 cpanel accounts. That's not true because each customer can create many accounts under their WHM.
Please show me where I said 40 accounts represented 40 Cpanel accounts?

Originally posted by freakysid
For someone who doesn't give a damn you have a lot to say! I am just pointing out that your Plan A and Plan B arguement doesn't at all gel with your business model. Your arguement is that plan B is the superior way to populate a server, yet because you offer multiple domain hosting, in reality your business model resembles something more like plan A.
The question was an example that 2 servers could have the exact same number of domains but have entirely different performance. I was not talking about our business model with that specific question.

Maybe you could answer another specific question -

Host A populates a server with 100 accounts. Each account has been promised 2Gb disk space and 10GB data transfer. CGI allowed.

Host B populates a server with 100 accounts. Each account has been promised 250MB disk space and 2GB data transfer. CGI not allowed.

Which Host's server has the best chance of being more reliable?

This question has never really been taken serious though has it Bob. :)

Gary

Originally posted by Editor
This question has never really been taken serious though has it Bob. :)
Can't say it bothers me much. :D

It's a dead simple question mate. Folks have their reasons for not answering it. ;)

You want to give the above question a shot? :)

Originally posted by Aussie Bob
You want to give the above question a shot? :)

No, because it is an unfounded question (with or without the cgi addon ;)).

Can you find an example of the scenario in play, and want an opinion on which is better? Or something at least close? Then I will. Otherwise, there's nothing to base the question or an answer on.

Cheers

Gary

Thanks Gary. Always true to form. :D ;)

BTW, I was clear that this was another question, so drop the "with or without the cgi addon ;)" routine. :emlaugh:
Maybe you could answer another specific question

Oh...just answer the darn question Bob! ;)

Can you find an example of the scenario in play, and want an opinion on which is better? Or something at least close?

Should be easy...unless the question is just too out there? :)

Gary

Originally posted by Editor
Oh...just answer the darn question Bob! ;)
hehe. A brilliant career awaits you in Politics mate. ;)
Should be easy...unless the question is just too out there? :)
There's no point Gary. You'll find something wrong with it and avoid it as per usual. :rolleyes:

I don't know what's wrong with this question -
Host A populates a server with 100 accounts. Each account has been promised 2Gb disk space and 10GB data transfer. CGI allowed.

Host B populates a server with 100 accounts. Each account has been promised 250MB disk space and 2GB data transfer. CGI not allowed.

Which Host's server has the best chance of being more reliable?

Originally posted by Aussie Bob
There's no point Gary. You'll find something wrong with it and avoid it as per usual. :rolleyes:

No, I give you my word. You want to keep pushing this question of yours, then simply justify it. And I will anwer it.

I do not avoid sensible questions Bob, you know that. Yours is far from sensible...unless I am overlooking something. It's only fair to expect your question relates to...something? :)

Show me a Host A and a Host B.

Cheers

Gary

Gary, there's no point. I won't play your game.

It's a hypothetical (http://dictionary.reference.com/search?q=hypothetical) question based on 2 exact servers that are populated differently. You claim that the question doesn't represent real life situations [what is reality?] with server population techniques and therefore you won't answer it. Who are you to set the conditions and terms of a question?

I don't think you grasp the concept of a hypothetical question -
hy·po·thet·i·cal ( P ) Pronunciation Key (hp-tht-kl) also hy·po·thet·ic (-thtk)
adj.
Of, relating to, or based on a hypothesis: a hypothetical situation. See Synonyms at theoretical.

Suppositional; uncertain. See Synonyms at supposed.
Conditional; contingent.

n.
A hypothetical circumstance, condition, scenario, or situation: OK, let's consider this possibility thenjust as a hypothetical.
The question I asked demonstrates that 2 servers can have the exact same number of domains, yet 1 server will have a better chance of better perfromance by what was promised to those domains. A no brainer there. :)
Host A populates a server with 100 accounts. Each account has been promised 2Gb disk space and 10GB data transfer. CGI allowed.

Host B populates a server with 100 accounts. Each account has been promised 250MB disk space and 2GB data transfer. CGI not allowed.

Which Host's server has the best chance of being more reliable?
Hence backing up the statement of -
The number of domains on a server is not always relevant to that server's actual performance.
Which became -
"The number of domains on a server is not always relevant to that server's actual performance. HOWEVER, the increase in the number of domains on a server will also increase the chances of someone affecting the server's actual performance."
If anyone can prove that statement wrong, post your arguments and reasonings here. :)

Originally posted by Aussie Bob
It's a hypothetical (http://dictionary.reference.com/search?q=hypothetical) question based on 2 exact servers that are populated differently.

It can't be tested or taken as true.

The only condition I have asked your question to meet is that it is sensible. It if can't be then why push for an answer? It's good to discuss these things Bob, but only if you're willing to learn. I am, if I am overlooking something. :)

Going back to where this all starts, where folks often come along and give some server specs and ask for some advice on how many sites they could consider putting on it. Wouldn't you agree the question should be:

Server A is a [specs] with 250 domains.

Server B is a [same specs] with 40 accounts allowed to host unlimited domains.

Both have the same combined diskspace and bandwidth allocations in total. Which server has the best chance of being more reliable?

I would say A. You?

Cheers

Gary

Nevermind Gary. If you won't show some common courtesy then how do you expect that to be recipricated? Thank you. :)

*shrug* I have no idea where you get that from Bob. But as I've said, if you can simply justify your questions, I'd be happy to answer them. You have yourself a nice day too. ;)

Gary

Originally posted by Editor
*shrug* I have no idea where you get that from Bob.
I love the "*shrug*" routine. Really Gary, it's almost too much. :D
But as I've said, if you can simply justify your questions, I'd be happy to answer them.
So the following question is not justified? -
Host A populates a server with 100 accounts. Each account has been promised 2Gb disk space and 10GB data transfer. CGI allowed.

Host B populates a server with 100 accounts. Each account has been promised 250MB disk space and 2GB data transfer. CGI not allowed.

Which Host's server has the best chance of being more reliable?
Please remember that the above question is un conjunction with the following statement -
"The number of domains on a server is not always relevant to that server's actual performance. HOWEVER, the increase in the number of domains on a server will also increase the chances of someone affecting the server's actual performance."
(1). How is this question unjustified in exact context of this discussion?

(2). Is the question out of context with the actual discussion??

Just to refresh your memory, here is the exact original context of this discussion -
Originally posted by WonderMonkey
How many "typical" sites can you count on per server? Of course I know you have to play it by ear and watch performance, etc. but there has to be a number you can use to estimate.
In that context I ask the above question and the above statement. Somehow you think the context of this discussion is about unlimited domain accounts. Oops. :D

But then again Gary, you are the master of taking threads totally off-topic :emlaugh:
Originally posted by Editor in this post (http://webhostingtalk.com/showthread.php?s=&threadid=89625#post698494)
In my spare time, I must admit that I specialise in taking forum posts off topic though. Please forgive me.
You're forgiven Gary. :D
You have yourself a nice day too. ;)
I always do. :cool::agree:

Jesues how many times is this going to be asked... Why don't we have someone merge them *all* and put it on a sticky and stop new ones from being made!:eek:

When you see a question you feel has been asked too many times, just stay out of it.

I was and still am fairly new here and am just trying to get a grip on things just like all the other first timers here. I use the search function and such but when it is not answered to my satisfaction I will ask it.

Its not my freaking fault that a few people can handle the simplicity of the question. For those of you who have and tried to answer, I sincerely thank you.

Oh how I wish, how I wish..... I could lock this thread now......

everyone in this thread needs a hug. :cartman:

WonderMonkey I don't mean to be rude, I'm just stating facts;). This has been asked at least 500 times, thus it should be stickied and making it all in one pot. :)
Originally posted by brian-WHT
everyone in this thread needs a hug. :cartman:
You can say that one again ;):D.

Originally posted by brian-WHT
everyone in this thread needs a hug. :cartman:
It's just a discussion. I like and respect Gary, but we're allowed to have differing opinions on certian issues. I don't see any flaming etc. :)

Mods can lock this thread if they must. I'm pretty much done anyways. :D

Originally posted by Aussie Bob
So the following question is not justified?

The penny drops. Let me know when you can justify it Bob. No time to go around in circles with it anymore. :)

Gary

Originally posted by Editor
The penny drops. Let me know when you can justify it Bob. No time to go around in circles with it anymore. :)
*shakes head in amazement*

It's totally justified Gary. If you can't show me where it's not justified, then that's quite OK. No need to reply. Let me know why you don't think it is justified in the original context of this discussion. Be accurate and specific please.

Here's the post again as you seemed to have completely missed it. :)
Originally posted by Editor
*shrug* I have no idea where you get that from Bob.
I love the "*shrug*" routine. Really Gary, it's almost too much. :D
But as I've said, if you can simply justify your questions, I'd be happy to answer them.
So the following question is not justified? -
Host A populates a server with 100 accounts. Each account has been promised 2Gb disk space and 10GB data transfer. CGI allowed.

Host B populates a server with 100 accounts. Each account has been promised 250MB disk space and 2GB data transfer. CGI not allowed.

Which Host's server has the best chance of being more reliable?
Please remember that the above question is un conjunction with the following statement -
"The number of domains on a server is not always relevant to that server's actual performance. HOWEVER, the increase in the number of domains on a server will also increase the chances of someone affecting the server's actual performance."
(1). How is this question unjustified in exact context of this discussion?

(2). Is the question out of context with the actual discussion??

Just to refresh your memory, here is the exact original context of this discussion -
Originally posted by WonderMonkey
How many "typical" sites can you count on per server? Of course I know you have to play it by ear and watch performance, etc. but there has to be a number you can use to estimate.
In that context I ask the above question and the above statement. Somehow you think the context of this discussion is about unlimited domain accounts. Oops. :D

But then again Gary, you are the master of taking threads totally off-topic :emlaugh:
Originally posted by Editor in this post (http://webhostingtalk.com/showthread.php?s=&threadid=89625#post698494)
In my spare time, I must admit that I specialise in taking forum posts off topic though. Please forgive me.
You're forgiven Gary. :D
You have yourself a nice day too. ;)
I always do. :cool::agree:

But I've said all that needs to be said. Not much more that can be added. What I've said still stands -
"The number of domains on a server is not always relevant to that server's actual performance. HOWEVER, the increase in the number of domains on a server will also increase the chances of someone affecting the server's actual performance."
Which is in perfect context with the original question -
How many "typical" sites can you count on per server? Of course I know you have to play it by ear and watch performance, etc. but there has to be a number you can use to estimate.
I'm damned if I'm going to continue this pointless discussion. I'm really bored with this Gary. Keep dancing if yous want, but I'm out. :D

Originally posted by Aussie Bob
I'm damned if I'm going to continue this pointless discussion.

Then stop editing your posts, and adding more darn questions until you can answer some yourself. :)

Gary