Re: [cobalt-users] [WANTED] RaQ2 Consulting

[Kris Dahl <krislists@xxxxxxxxxxxxx>]

on 9/1/00 10:28 AM, Rodolfo Paiz at rpaiz@xxxxxxxxxxxxxx wrote:

>> If you don't have a much time or money, I would highly recommend
>> reconsidering your business plan.  Do you have the skills and
>> the funding necessary to run a successful hosting company?  There
>> is very little room (IMHO) for low-end providers, as there can
>> only be one least cost provider--and usually that is someone with
>> lots of VC capital to burn.
>> 
>> You may still be able to make a go at it, but I would make
>> very sure that it fits in with your core business.
> 
> Actually, once you leave the USA, the market changes somewhat. In
> Central and South America, for example, bandwidth tends to get
> rather expensive (read, up to $25,000 per month for a satellite E1
> or $20,000 per month for a T1). So, many times international
> bandwidth is somewhat restricted since providers buy only as much
> as they need. The predominant reliance on satellite links also
> introduces 500-600 ms worth of latency which makes quick web
> downloads rather more difficult for the average customer.
> 
> For the Central American markets, I estimate that we would only
> face four or five significant competitors in hosting, and by
> "significant" I mean with 100-200 customers each.
> 
> I've also redefined "low-end" to mean two new RaQ 3i with 128 MB
> of RAM. Of course, I do have redundant links to the Internet and
> a 2 KVa UPS... so although I'm definitely low-end and small, I
> should be able to offer at least a 99% uptime guarantee while
> reducing costs since the site and most of the browsers are local
> and thus don't incur satellite costs.

I agree that the market is different in other places--and I refer to you for
that experience.  It actually sounds somewhat like what the market was here
5 years ago.

With the lack of competition and a high cost of entry, if the demand is
there (and as costs decrease) there will be a big opportunity for you.

>> Well, it will actually offer very little redundancy in and of
>> itself.  It would require a minimum of effort to have them back
>> each other up for email, dns, but load balancing for web is a bit
>> trickier.  Do you have plan for how you are going to do this?
> 
> No, but I'm going to have to make one, I know. I'm considering
> StaQware from Cobalt (you'll notice a post to the list asking for
> opinions on it), and I'll need to learn a lot before we go live.
> 
> Any suggestions in this area, Kris?

I saw that post... I am not really familiar with the product (although I
probably should be).  It looks interesting from what I am reading.
Certainly going to be cost effective.

One concern is that you get no benefit from the second Raq3i until you have
a hardware failure.  Which kinda sucks, IMHO.  We use Linux High
Availability (HA) with Heartbeat.  We load shape with round robin DNS and if
one fails, the other automagically takes over the other's load.  Pretty hip,
actually, and works well.  Both servers communicate with a database server
that has Raid, etc.  So we get the high availability *and* increased
performance.  This is not a Cobalt solution, and is not running on Cobalt
gear.  It is running on VA Linux gear and I set it all up by hand.  Took me
about 3-4 days to get all three servers ready for productions--so that
represents about a $3000 investment time-wise--much more expensive than
Staqware.

Also we are running only open source software on our gear, for security,
etc., reasons.  I don't think Staqware is OS.

Why is it that Cobalt will use other people's open source stuff yet isn't
releasing any themselves?  This is one of the main reasons we went with VA
as a vendor (besides the lack of high end servers from Cobalt) over all
other choices.

>> I don't want to get you down at all, but running a hosting
>> company is not easy or cheap--to do it correctly.  I just
>> want to warn you ahead of time of the costs, time, knowledge
>> that are involved.
> 
> Thanks; the intent was clear. However, I have bandwidth, rack,
> and UPS capacity already. I need to buy two RaQ's at least to
> get started, and I have two friends who run hosting companies
> in Mexico who are willing to help me get off the ground and
> don't mind me competing. They will admin the boxes for $6 per
> month per site, which is fine by me. Everything else I get to
> keep. I will happily reinvest all profits to improvement and
> growth; then, if I can manage to put 100 sites on each RaQ and
> get each box to generate $2,500 per month, I'll revise the
> operating model for the business. If it's a bad idea, I'll
> have to sell the RaQ's, and give the customers to my friends.

Excellent.  I am glad that you have a support net... waaaay to many people
just jump into this.  By doing your due diligence and planning for things
like this ahead of time you are saving yourself some serious headaches and
insuring your success.  I applaud you for doing this.

And I think their support cost is *very* competitive and is going to save
your ass more than once.  Its a small price to pay, I guess I am saying.

Slighly OT: You ever mention to someone that you are in the Internet
business?  I have heard several times in response: "Oh yeah?  I have a
server."  I press for more details and get something like this: "Yeah, its
Windows 2000 Box running Microsoft web server [i.e. not using proper
terminology].  I'm hosting a couple sites on it off my DSL line."  It bugs
me because it seems like everyone is offering 'web hosting', yet few do it
right.  How many people are happy with their current web hosting company?
Very few, right?  It gives professionals like me a bad name and waters down
the perception of the industry.  So I am all the more happy to see someone
take it seriously.

I wish you luck.

Cheers,
-k