Hello again everyone. Gosh, I must say..I AM LOVING this forum. It's such a quick answer to anything!

I am considering starting a web hosting business. I HAVE a considerably large sum of money that I plan on investing into the business, but before I do so...I want to make sure I choose the correct route.

Now, I see companies (Example: ThePlanet or EV1) that are actual datacenters who have HANDS ON access to the servers and then I see business model that offer hosting, claim to have datacenters "remotely" and they obviously charge way less.

So if I want to start my own datacenter, and decide to invest my money into the infrastructure required, do I just need to buy:

1. Needed Software
2. IP's
3. Pipeline & Incoming Cable Infrastructure
4. Servers
5. Develop website

Please let me know if I am missing anything. This is something that I seriously am considering doing... also, how often do hosting companies start-up that ACTUALLY PHYSICALLY OFFER HOSTING (not resellers)?

<3

-dr00t

If you honestly are thinking about doing your own datacenter and you can't look at your bank account and see $1,000,000.00+ USD sitting there ready to be liquidated on a chance, stop thinking about it and start thinking about co-location if you really want "hands on".

Well, I certainly wouldn't purchase as many servers as ThePlanet has immediately off the bat...that would be just plain stupid.

I would buy them in sets of 20 or 30... so I would start off with 30, and as soon as they start to get full, I would buy 20 more....and so on and so on and so on...

Then why not co-locate?

1. Needed Software
2. IP's
3. Pipeline & Incoming Cable Infrastructure
4. Servers
5. Develop website

While I guess these are the core neccesities to host your own servers, I dont know if this classifies as a "datacenter." Maybe its just me but when I hear the word datacenter, I think of unbelievably large amounts of bandwidth directly from internet backbones, overhead fiberoptic raceways, expensive backup power systems from ups's to diesel powered generators, powerful and reliable a/c units all inside of a building secured with the latest in technical security systems. The cost of the servers being nothing compared to the price of setting up the required storage environment.

I dunno, maybe I just been looking at too many datacenter pictures on web host sites :P .

Well said by janton. I think the environment of the datacentres will cost more than all the servers and softwares. Because you need to ensure a secured environment for your datacentres.

Try looking at Carrier Hotels before you jump into building and maintaining your own DC. www.carrierhotels.com

There is alot more then just what you have listed, and I would recommend you learn your way around the industry before attempting to even get into a carrier hotel or colocation for that matter. Not to be rude by any means, however comparing owning and operating a datacenter to reseller hosting is just not practical. Start off small. Crawl before you walk, and walk before you run. Your running before you know how to crawl.

Perhaps try running a single server first on a very small scale and see where you get with that. Alternatively perhaps try and invest into an existing data centre if such a thing is possible (or offer to rent servers from a data centre).

As others have said it does seem that running such a centre is pretty costly (the physical cost of the equipment, renting suitable places to house the servers, managing them, enforcing security and preventing against things like fire for example) and most business venture require a good knowledge and fundamental understanding as well as money to succeed.

Thank you all very much for this information!

I saved up quite a large sum of money after selling a big advertising agency based in Seattle. But now, I am looking for something to do to fill my time.

Hosting is something I think would be fun, and I do in fact have quite a bit of knowledge with networking... However, I do plan on starting small..and with time, the datacenter thing IS something I want to do.

This forum is a MAJOR HELP, without a doubt.

Thank you all very much. The investment idea sounds very lucrative as well.

Once again, thank you. Now brace yourselves for more of my newbie questions!

-dr00t

Ahh, I have a friend here locally who wants me to add another question...

As I warned, please brace!
---

He wants to know... would it be a good idea to just purchase T3 or OC services and begin renting a small office somewhere and providing hosting via the T3/OC Connection? What are your thoughts on this? Wouldn't this be a way of saving money on Co-Los and the whole datacenter thing?

Thanks again!

Again you are really jumping head first into something you're not ready for. Are you sitting on $1,000,000+ ready to invest and possibly loose very quickly for this project? if you're not, owning your own datacenter is a horrible route to go. Look at co-location, possibly your own private cage/suite within an existing datacenter, or anything along those lines.

What are you acting like you want to do is just a step above a simply "basement datacenter" idea.

Why do you feel the need to open up a datacenter over the other options out there?

Thanks again for the quick response.

You see, I KNOW I am ahead of myself, so you need not point that out anymore. I have acknowledged that fact. Which happens to be the reason im here!

You ask how much I am sitting on ($)... Yes, well over the amount you estimated. HOWEVER, that doesn't mean I won't do OTHER stuff with it as well.

I have already set aside much of the more important and critical monetary responsibilities needed (Family, etc.).

Let me give you more insight about my ambitions...

I am an investor. Investors invest in things they feel will bring a return or offer futures, jobs, and make a difference. I am not the style of investor that typically settles for a half-assed idea or goal...I like to do things RIGHT from the START.

I am simply hunting down the right people to learn from (YOU ALL) and laying the groundwork for something that will eventually be run and managed by people that I have little or no control over. Only to be re-bought or aquired by someone larger. In the meantime, I want to provide the excellent service known to exist and be competitive, but that is not what I am asking about at this point. I want to learn about the tasks at hand, which are UNDERSTANDING the idea to begin with.

Hopefully this helps you, help me a little better. Cut out the lectures and give me the good stuff! Thanks for your input so far.

-dr00t

Well in all honesty this is a great place for a few starter question, but a professional consultation will do alot more for you than we can. It's nice to hear that somebody actualyl has the money for once as compared to the general:

I have $5000 this will get me a dual xeon machine, router, and a cable modem right?

As an investment a datacenter would be rated quite poorly actually; there are many different investment oppertunities out there that will bring a much larger return and usually in a much quicker manner than a datacenter. It would be different if we were talking about you already having a large established hosting business that you were wanting to open up a datacenter to cater to new clients with.

It's hard enough for a dedicated provider to start up in an existing location, let alone somebody building a completely new datacenter out there. When I can get a machine with good hardware already within Savvis for $65/mo, how are you going to compete with that for example?

That aside as I know you're getting sick of my anti-build your own datacenter posts ;) It's best to start with a general list:

1. Building (yup, you can't put that equipment out on the street).

1a. Cooling, believe me servers get hot (talking about a single rack even) and they will heat the heck out of a room in a hurry. AC is going to be a very big conecrn. High Heat + Server = Meltation (not sure that's a word, you get the idea though.

1b. Power: Servers use a lot of power and your AC units are going to use up quite a bit as well, keep this in mind when choosing a building.

1c. Along with power, backup power. You are going to be needing a good UPS system + generator switchover most likely (at least if you want anybody to host in your datacenter you should be thinking about these things).

1d. Often overlooked but: Expandiblity, moving a datacenter is not fun for you or your clients, having your growth stunted while you hunt for a second facility plus the costs of managing a second facility are not going to be fun either.

1e. Fire protection: You are not going to want sprinklers raining down on millions of dollars worth of electrical eqipment, and nice fire supression is going to cost you some money.

2. Infastructure (this is closely related to the building). That beautiful farm house in the middle of the country sure would be nice for a summer vacation, but it is not going to be great for getting proper internet access out there.

2a. The actual connections. You were speaking about if you should start with a single T3 line, OC based line, etc... Let me ask you this: If you were looking to buy a server (exception to the rule is highly used unmetered machines running pure cogent) would you want to know you're connected to a single provdier and limited to a max of 45mbit? In fact the entire datacenter is limited to that amount of transfer? Start thinking mutliple connections to multiple providers and the term everybody loves to hear: BGP

2b. Routing those connections. Getting internet from your main pipes in over to your servers doesn't happen through thin air (unless of course you are using wifi) and even then you are still going to need routing equipment. You say you want to start small, but upgrading a core router later on because you skimped on max compacity is not going to be fun.

2c. Switches. You are going to need them and I'd be looking at higher end switches allowing you to do remote management and monitoring.

2d. DDoS Protection (Not 100% needed, but another thing to look into). Tipping point and Fireslayer come to mind off the top of my head.

2e. Cables: There will be a lot of cables run throughout the entire building, you are definatley going to need some way to manage all these cables (raised floor comes to mind in a hurry here) and such a structure is not usually found in an building and instead must be built by you.

3. Harware, the meat and potatoes of your datacenter.

3a. Just to mention from above, routers, switches, cables, patch panels, etc.. oh my!

3b. Racks/Cabinets you are definatley going to want these (including bakers racks most likely).

3c. Can't forget the servers!

4. Security

4a. General security of the structure such as cameras and locks on every door.

4b. Security based staff (or regular staff that also handles security) on site 24/7/365 to handle any needs of customers.

4c. Card readers, badges, palm readers, etc.. are all options you should be exploring.

5. The best category of them all and often forgot about, Staff!

5a. If you want to have a good datacenter, you are going to need staff and they are going to need to be there 24 hours per day. Assuming only 3 people there at any given time, 24 hours per day, 7 days per week, we're looking at 504 hours a week that you are going to have to pay out. Now pretending you have a stupid set of people working there for a tiny $10/hour you are looking at $5,040 per week (not taking into account the taxes that you are going ot have to pay on that). 52 weeks in the year we're talking about a running cost of a set of staff that will not keep your datacenter running at $262,080.00 per year

5b. Great Staff! Not sure where you're looking to open up at, but $10/hour is not going to find you any top IT type guys that will be able to keep things running all the time. You are going to be looking at even more to get one nice senior network administrator in there


The list could go on and on, but that should at least give you something to chew over.

Interesting.

Now that we have gotten beyond the money concerns, we are getting somewhere!

Great post! Must have taken a shitload of time and thought. This post will be referenced a lot, I assure you.

First of all, this is a long sought plan in the works that you are contributing to, and professional counseling has been another piece of the pie. However, I like to look like I know what I am talking about when we are at the meetings with investors and consultants, therefore I come here on my own time. Haha.

Shh, don't tell anyone my sources though!

Did you ever step back and realize how lucrative the "NON-Hosting Business Model" of this Datacenter ordeal really is? If not, maybe this article (http://www.cnn.com/2006/TECH/03/28/rural.techboom.ap/index.html) will help you.

The internet is in as high of demand as oil, and with older generations (non-net users) decreasing and new generations using the internet...there will only be an increased demand for all of this content and information we need to satisfy our everyday Google searches.

Hopefully this perspective helps you realize the potential of my plan.

Also, when you start adding up numbers...(which are very low considering you didn't add the power bill.) you forgot do decrease monthly revenue from the amount. If the business is operating at full capacity and costing a buttload to run, then it is obviously doing very well and if the cards are played correctly...it should pay for itself to run AND turn a profit (at some point).

Once again, you have far exceeded any expectations I had for any information on this website. The consultants will **** themselves when they find what I have learned on my own! We will keep this our little secret.

Cheers,
dr00t

Took a decent amount of time and thought; however, I should have prefaced it with I am not the very best person to take advice from. When it comes to the ins-and-outs of a datacenter there are quite a few members here more knowledgeable than I. I do have the benefit of offering a customer perspective though, and that should be worth at least a little bit ;)

The article you are posting is quite interesting, but unless you have other operations that need insane amounts of diskspace it really would not be worth the money. The companies listed in the article are looking to open up data storage centers for their Search Engines which makes it a very worthwhile investment to them. The more data they can store the better their services will be and more customers will flock to their already wonderful (sometimes at least) service. Take a look at each of the companies listed there and see where they started from though, none of them started out building a huge business.

Google was a college project IIRC
Microsoft was partially purchased from another person and attempted to be marketed to Mac at first (somebody will correct me on this one, my memory is hazy)
Yahoo escapes my mind at the moment.

The internet is most certainly in high demand at the moment, but the demand aslo comes with certain expectations of what a datacenter should have for the most part. Hosting is part of a datacenter 90% of the time I would estimate (the only thing I can think of non-hosting related would be to simply use it as a routing point for lack of a proper term).

The numbers I tossed out there were quite low, the general rule of thumb is $1,000,000 or so to sink into it you could get started with something small and really hope for the best. If you want something halfway decent start thinking along the lines of $10,000,000.00 of course. That number will definatley change quite a bit depending on all the different variables (are you moving into an old DC with cooling systems in placed, rasied floors, etc.., how far are you from different providers to get your lines in there, what level of staff do you want, power costs in different areas, how large of a commit you want to start with, pricing you can get on your servers, and on and on and on and on).

Now another item you should really be looking at from the very start is what type of target market you are planning to aim your services to. RackSpace has a much different network/datacenter setup than you are going to find at gigeservers and still different yet from the one LT runs from within the Savvis datacenter.

IIRC gigeservers runs some of the best DDoS migration stuff out there and they've been able to handle multiple gig attatcks with no problems at all, if I was to start pulling a 5gbit attack on one of my machines over at LT I am quite sure they'd pull the plug on that box (and it would be something I'd expect them to do from the start). If I am hosting with RackSpace I generally have something mission critical. I do not want to find out htey only have two Gig-e's (one from cogent one from level 3 for example) and I don't want to hear about a single core router failing on them, backup generators not comming on, etc.. Last time I spoke with the people there they have had 0 downtime in the last 5 years (best in the world as far as a datacenter goes). They can obviously ask and get more for their product than you LayeredTech could, and both services appeal to a completely different customer group.

One more point to add as I'm thinking about it is a proper growth plan. Having been watching many companies grow/mature for several years (and in addition reading threads on providers here at WHT) I have seen that many great companies simply can't handle the growth when it starts to come. This applies to dedicated providers, shared, VPS, and just about everything else. There is a recent thread about The Planet and their growing pains that might be of a decent read to you (if you can't find it ask and I will post the link). If you don't have a proper growth plan in place you can really get hit hard down the road.

If you think of any more questions please feel free to ask. I do appologize for getting on you about having the capital for doing a setup like this, but it is common for a person to come here to WHT asking how to start a datacenter and not having much money/experience/etc... at all.

"I just need a few servers, a switch, and a firewall right?" <-- Typical train of thought of most users here.

I also don't want to drive you away from WHT, but a great supplement to your reasearch could be: http://datacentertalk.com/

Firstly droot, dont take my post as an insult, more as trying to be blunt and helpful.....


After reading another topic youve started:

http://webhostingtalk.com/showthread.php?t=499591


I am a complete newbie to the hosting scene. In fact, I have little to NO knowledge of what is going on. I don't run a website or anything fancy like that, but I do want to know a few things before I even consider any of that.

My question is...

What is the difference between ThePlanets "Total Control" services versus Site5's services (reseller). I mean, the only thing I can tell is pricing...however, there HAS to be something different about them for the difference in price I see.

I don't even know what the difference is between the two business models, please enlighten.

-dr00t


I wouldnt recommend jumping in at the deep end of servers and colocation, id advise getting a resellers account with a reputable provider for a while and get yourself some "hands on" experience with the way various software and systems work.

From your other post it is clear that you are not certain on the difference between resellers and dedicated providers, sure the end result (website) is the same but the behind the scenes work needed and experience needed varies greatly.

Again, not intended to offend you but more a sensible suggestion from someone whos worked his way up from shared hosting.

Good luck with whatever path you choose to take.

Moved to Running a Web Hosting Business ;)

In all honesty, if this is a serious proposal then I don't think the OP needs to know how to run the day to day aspects of a hosting business.

A serious investor with a few million $$ will employ other people to do the specifics, but would quite rightly want to do their own investigation into a possible investment industry.

When talking to consultants and potential directors/managers and employees then you'd want to know an overview of the industry and practices so you do have an idea of what is going on. Otherwise the people you are relying on could be giving you bad advice that you have no clue about whatsoever. At least with your own knowledge you can ask meaningful questions in your meetings and get useful information back. You still have to start from the real basics though.

As an example, do you think Richard Branson knows every little thing their is to running a mobile phone company or air travel business? I very much doubt it. He looks for industries to invest in and then relies on others to carry out his plans. That's not to say he doesn't do his own research to get as good a general overview as possible before he starts talking to consultants etc.

If the OP really does have the financial backing they claim to have then this train of thought does make some kind of sense to me. I could quite believe someone to have an idea of an industry to invest in but have no practical knowledge of that industry.

mjb-is:

The best way for an investor to be overrun or lose money is by not knowing about what they are investing in. That is why there are specialty investment or venture capital firms (Ex. tech investors, agriculture investors, etc.)

For the rest of is individual investors, we have to take a hands on approach.

Thanks for your input!

-dr00t