Market for a Secure Data Center?

[Brian S]

Hey folks,

Just a quick posting to receive some feedback. I'm looking for your thoughts on a theoretical situation, as well as the market for Secure Data Center services in general.

First, what are your observations on the market? Do you think this is a segment in demand; is it a profitable niche?

To be frank, this not an area I have a great deal of knowledge about. If you have suggestions for other sources of info., I'm all ears. Here's why I'm interested...

Let's says there's this large building -- a former military secure operations center. I need to find out some more info on its specs, but as I currently understand, it's made to survive a non-direct explosion. With some very thick concrete walls and a secure operations center, it's built to withstand damage from the outside.

Other potential benefits:

- Located in a geographically unpopulated area (safe)
- Wages are low
- No major geographical disasters (tornadoes are very rare, snow is the biggest "threat")
- Building is a equipped with a large generator building
- Could already be partially constructed for datacenter operations.

Possible drawbacks:

- Located in a geographically unpopulated area (7 hours from a major city -- travel may be inconvenient.)
- High bandwidth is not readily available and cheap.

Here are my preliminary thoughts: The business proposition is to enter a possibly more profitable secure data center market. Marketing would place a heavy emphasis on the security of the building, with services appropriately marketed and priced to businesses looking to host in a secure building. If the datacenter could be had with a reasonable investment (there are a number of financial incentives being offered), it could help offset the costs of bringing in redundant bandwidth (e.g. a T3 from ATT and Sprint). The building itself is up to date as of the mid nineties, so the physical aspect of the building is solid. It is anticipated, with what I know about this building, that it may already have some of the trimmings that make a datacenter (raised floorings, etc.).

I wanted to get some outside input and advice before I research this further. I would love to hear comments from anyone who has first-hand experience with datacenter operations. The way I see, it could be viable proposition if: 1.) There's a growing market for secure datacenter hosting. 2.) The datacenter could be obtained and furbished for a reasonable sum. 3.) Redundant fiber could be brought to the building from two sources for a realistic sum (it's about 150 miles from a major Sprint backbone, and 5-7 hours from other major fiber lines -- a long way).

So what do you think? Is it possible? Have companies like USDCO ( http://www.usdco.com ) tapped the market?

Thanks,

Brian

[Rob-Bocacom]

Hi Brian,

I am by no stretch of the imagination an authority on the secure data center market in the sense that you are using “secure”. So, I’m not sure about that question. However, I am pretty knowledgeable on what it takes to build out a small facility into a first class data center. We recently took a similar situation – partially outfitted space, middle of a concrete building, generator farm, etc and launched a dedicated server business. We ran fiber across and down a hall a few hundred feet to connect to a major backbone. The experience was great. It was fun, it was cool, it was exciting, it was really hard work. But if I had to sum it up in one word, it was expensive, and continues to be expensive every month.

We have a very secure facility, in a very secure building. Across the hall is Verio. They took the same secure building and added a concrete bunker around their datacenter. They charge a lot for colocation, but they are probably able to satisfy most of the same customers you would be competing for.

So, my thoughts on your idea are that there would be better options to invest money in a business. You mentioned there are incentives available, maybe there are enough incentives to offset the costs, but my instinct says probably not. There may be a market out there, but I think it would be very difficult to succeed without a large amount of capital, especially if you don’t really have the background for it.

But you are doing the right thing by researching and I’m sure you’ll learn a lot in the process – always a good thing. Do a preliminary business plan and figure out if it’s a viable plan. I would start by figuring out the costs as you’ve got some big ones to overcome… fiber, equipment, marketing…

With all that said, I would be interested in hearing more about what you learn and if you decide to pursue it. Keep us posted.

[tranz]

ok, here it goes
Yes the facility would be a good one IF it were located a hell of a lot closer then the distancees than you are describing.

The fiber runs would have to be boosted over that length and optical multiplexers installed.

By the time you are done the cost would be so high that you would be unable to get back your investment anytime soon.

I mean we are talking millions just to get the fiber there in the first place, much less the building.

What you have to ask yourself is "Why hasnt anybody else been there" COST..

Also ask yourself why most data centers are located in the center or as close as possible to populated areas. Because the cost of deploying the needed infrastucture is minimized. The rings and, or backbones are already there.

The data center that I am in spent $14.8M in order to have it built out the way they needed it. It is located in a major populated area of town across the street from major fiber backbones coming through the area.

I have just been talking about running your connections. You have to think about power, yes there is a generator there but you have to bring in Dirty city power, clean it up, send it to your HUGE UPS system and then onto you data center floor.

There are a huge number of factors that have to be thought about and it just isnt cost effective anymore. During the height of the dot com days when people were spending millions upon millions of venture capital money they went out and built facilities that you are speaking of. Some to the tune of $30M to $45M.

Almost all of those facilities were sold and for pennies on the dollar.

Large enterprises that need that type of security in most cases will keep their gear in house. Why? There are to many people where their jobs and careers are on the line if something happens. They want it fixed now and to have more control. Basicly, they dont trust anybody else to do it the way they would. Even though the way they are doing it might be wrong.

Smaller guys will put the gear in locations like Rackshack, ThePlanet, Rackspace and me.

They dont need the nuclear hardened facility. if it works and it does have security to stop someone from going crazy at 3am by getting in and damaging their gear then they are happy.

The market isnt there to support it. Some though that it would be but it isnt.

Hope I answered some of your questions.

[carrierhotels]

Brian,

Don't even think about anything like this unless you have a ton of money that is a completely speculative capital. USDCO is one of a number of well-funded operations that have positioned themselves in the ultra-secure hosting market and are already up and running. Unless it's plug-n-play for advanced hosting (which I doubt), it'll cost more than you think to fit out.

The remote location is good in theory, but bad in terms of actual everyday business practice. Fiber providers won't light the building unless you already have enough business to cost-justify their expense. Power is a huge issue nowadays. Advanced apps require more power, which equals more cost.

But most importantly, enterprise companies looking for ultra-secure backup need to be able to get their IT employees to the facility quickly in the event of a disaster. That's difficult in the wilderness scenario.

Instead of remote locations, many huge financial companies are buying backup data centers in the suburbs of metros that are probably low on terrorist target lists but have good transportation to major hubs. The Kansas City, Dallas and Atlanta markets are all seeing excess data centers bought up. Our web sites have more on this, as we cover this market full-time.

Rich

[allan]

Take a look at http://www.servervault.com/ they have implemented a secure data center, with an infrastructure similar to what you are describing. I have toured their data center, twice, and it is pretty impressive. However, they have already filed chapter 11 once.

They emerged a much leaner, and more frugal company, and have been successful largely because of several big government contracts. But without those contracts, I am not sure they would be able to survive the long haul.

[Brian S]

Thanks for the feedback guys. It is what I was looking for. I'm going to continue research this and I'll let you know if anything comes of this idea. Even if this venture isn't feasable, something else may come of it. Sometimes with bad ideas, come great ones.

Thanks,

Brian

[DizixCom]

Sorry to revive an old thread, but I thought I should make a note here.

USDCO is no more. I was with them for 1.5 years and loved it. It was an excellent facility, an exciting place to go. But once you get past the novelty, it was really just another data center. Always interesting to get on an elevator at ground level, then go down 80 feet.

[allan]

Originally posted by DizixCom:

Sorry to revive an old thread, but I thought I should make a note here.

USDCO is no more. I was with them for 1.5 years and loved it. It was an excellent facility, an exciting place to go. But once you get past the novelty, it was really just another data center. Always interesting to get on an elevator at ground level, then go down 80 feet.

Yea, I had heard that -- and HavenCo is having problems as well :(.

[DizixCom]

Ick. That's the Sealand facility out on the old oil rig? It'll be a real shame to see that one go.

I'm just speculating here, but I imagine the market for such secure colocation is nearly non-existent. A good insurance policy is likely a more cost effective way to deal with disasters.

[allan]

Originally posted by DizixCom:

Ick. That's the Sealand facility out on the old oil rig? It'll be a real shame to see that one go.

I'm just speculating here, but I imagine the market for such secure colocation is nearly non-existent. A good insurance policy is likely a more cost effective way to deal with disasters.

Yea, that's the Sealand facility.

I agree with your analysis, except that it seems like a combination of good insurance and good backup is what businesses are after, because Iron Mountain (same business model as USDCO but they only kept backups in their mountain) has had two very strong years.

[Adam]

I think it's a good idea above all esle and as previously mentioned but yeah the major factor is cost. You'd need at least $1b to secure yourself stably for a while is my guess.

I personally provide venture capital etc. and as attractive as a prospect like tihs would look i would decline it because of the amount of time i'd have to wait to get the money back (without the APR) and how unstable the investment would be.

Just a few weeks back i turned down one data center proposal, and that was just a standard data center/noc.

It's good you are researching it all and thats commendable however i think you should look into something smaller to start with.

Companies such as BurstNET and JagurPC were all very much smaller for a long while before they eventually migrated to their own building which at that stage was still classed as a risk for them back in the day im sure.

If you're intent on an opening of an entirely new building and isntead of making the focal point of your data center how secure it is, why not make the focal point of how reliable it is using only the best equipment etc. Yes its lots of money again, however its a one off cost and then you're set and the only other costs you have is the essentials...power/bw etc.

I hope this post was able to help you in some way or another.

[Brian S]

Shame about USDCO. I shelved this idea not long after this discussion. The project was far too big for me to personally handle anyway. I was the visionary. Main reason being the high cost for bandwidth, not to mention the build out for a facility that was only partially outfitted for a server environment. Add to that my little experience in this realm, and it didn't make sense. I do think it was worthy of chatting about. And that, we did. :)

Brian

[Adam]

It certainly raised a lot of important issues and sub-topics so yeah this thread was well appreciated by me for sure and im also sure that all people share this appreciation :)