Slashdot Mirror
Starting Up A Colocation Service?
mhoskins asks: "I am considering the possibility of opening a Colocation service. I am trying to gauge the need amongst the community. Is there still a large need for low cost hosting for LUG's and projects? What would be acceptable pricing? Are people looking to host their own boxes or something like the CobaltRaq? What are the likelihoods that equipment will be bought, leased or rented? We expect of offer advanced services like host monitoring for those who want it. Is there any interest in that kind of service also?" Colocation services might be better for proto-Internet Businesses than spending the (large) amount of money needed to get wired.
Location is what most projects are lacking, whether it be co-location or project-hosting, or whatever. I like the idea of being able to bring in a box or two of my own, use your rack space, your network drop, and your power to power my box(en). In return, I pay for bandwidth, rack space, and power consumption; according to usage for bandwidth, and fixed rates for the latter two.
For this kind of service, basically you need a big fat pipe coming into your space (probably more than one is ideal, each attached to different backbones), a lot of power, and space. Co-location facilities often have cages or some other type of physical protection for the boxen themselves, but you could probably knock off some of the space charges if customers bring in their own cages.
I would say you will get mainly bigger setups as customers. Smaller businesses/projects tend to be OK with just web-hosting, rather than full-scale co-location (I may be wrong about this). Note than I mean big in terms of both number of boxen and amount of bandwidth.
Services such as DNS and the like are unrelated to co-location -- if you put your own boxen somewhere you should have one to handle DNS requests as well. I would concentrate on the power, space, and bandwidth issues and leave the extra services to the customers (or at least put off offering them for a long time).
Good luck!
darren
Cthulhu for President!
(darren)
This is what i want from a coloservice.
1. multiple backbones
2. shaped rather than metered service (ie if i pay for 384kbps, i want you to enforce that w/ traffic shaping, rather than billing me more cause i went over)
3. safe storage for my box(en), i don't want them getting burned, earthquaked, flooded, molested by other customers, etc
4. visiting hours, i want to see my box(en) cause it(they) miss seeing me in person
5. UPS, and not the wimpy ones that are just the size of a full tower, i want the whole building to have a diesel gen for power outs
6. Location, it needs to be close to me (duh)
7. a decent web page, when i go looking around for colo its a pain in the ass cause ppl don't put their prices or their backbone providers or something is always missing
if you're running it out of your garage, i'd rather run my boxen out of my garage than pay you to run them out of yours
Need a Catering Connection
I don't think that this type of thing is needed anymore for free software projects (or maybe even LUGs) now that SourceForge is available. It's free, supports PHP3 and CGI, provides free forums, mailing lists, FTP, CVS, and other stuff. This is more than enough for most projects.
Well, if you're wanting to ensure uptime and connectivity, you definitely won't want to use broadband. They're absolutely great for personal use, but in the event of an outage, they're relatively low-priority (behind leased lines like T-1's, etc.), so a co-lo service would not be a bad idea, assuming they're multi-homed, etc.
"You can never have too many elephants on your team."
Your target market depends on your size. If you are thinking of running a T line or two into your garage, you arent going to get taken seriously by many business customers. If you are looking to lease a building that is sitting atop a fiber backbone, you need to have your business plan done by MBAs and not Slashdot. What you do and who you can target will depend on how much money you have to get started.
If I was going to do this, I would combine colocation and consulting in one business. Then I would target small businesses with Internet needs and few tech workers. Check out real estate companies. They loathe hiring full time staff who arent making sales. Generally are slow (but interested) when it comes to picking up new technology. However their business can really benefit from a strong Internet presence. Show them the benefits, especially the ability to access data from home (agents love that), and then put together a package for them. Once you get the contract, you get monthly income from the hosting, and you will be their consultant because you manage their server. This wouldnt be too expensive to setup because you wouldnt need much office space, and your bandwidth needs would grow much more slowly than a "colo for 50 bucks a month" deal.
OTOH it is a lot more work than running around hitting the reset button and popping out floppies. However I do not see a rosy future for low end colo-only companies. What happens when having a big, reliable pipe to your home becomes the norm? You know when that happens, it will be cheap enough for businesess to have mega connectivity on their own. When Joe Dentist has as much bandwidth as you do, what vital services are you offering them? If they dont need you for something else, no one is going to want to stick a box in your rack. On the high end, people sitting on a huge bundle of fiber and selling hundred megabit plus connectivity packages are still going to be doing OK. Can you get there in time? Maybe you should you focus on something else and use colo as one of many products you offer.
Good luck...
-BW