Linux on a Fast-Food Diet
Aug 23, 2005, 23:30 (6 Talkback[s])
(Other stories by Brandioch Conner)
Re-Imagining Linux Platforms to Meet the Needs of Cloud Service Providers
By Brandioch Conner
Instead of the traditional car analogy, how about a restaurant
At one end, you have yourself. You go to the market and get the
basic ingredients, mix them yourself, cook them yourself in your
own oven, and eat them yourself. This is equivalent to downloading
an Open/Free distribution and sorting out what you want and making
it work the way you want. Most of you have done this (I'm not
counting desktops, yet).
At the other end, you have the five-star restaurant. Everything
is handled for you, for a price. This is the equivalent of
deploying Oracle or such. A few of you have done this.
But what about the middle? Where's the "McDonald's" model? By
that, I mean a limited selection prepared in a limited number of
variations but delivered to you quickly and inexpensively. Sure, a
meal at McDonald's might not cost a tenth as much as a meal at a
five-star restaurant, but McDonald's (the corporation) makes
billions of dollars a year, and they're everywhere.
Linux-based companies will never see the same profit margins as
Microsoft. They're supposed to be making all their money on
"service" and "support" and "customization." So, what about the
contractors? They'd be the ones doing the deployments and providing
the support in the "McDonald's" model.
Here is my breakdown of the various roles possible in a
- Vendor/Salesperson (example: Red Hat/Red Hat salesperson)
- Corporate executive (CxO or owner)
- Tech/admin/contractor/consultant (on staff or contract, maybe
not even on site)
- End users
There are a lot of small companies that don't have an IT person
on staff. These are the perfect organizations for basic Linux
services, as long as they're inexpensive, reliable and
extensionable. All it takes is looking at the roles
(vendor/exec/consultant/users) and determining their objectives and
then fitting a service to them at a price point they're happy with.
Sell them the service and support, don't sell them "Linux."
For example, Groundwork has a really nice
product and it's Open Source. But why aren't we seeing Red Hat or
Novell offering very basic monitoring functionality for $10/month
per 10 servers? Novell could setup the monitoring system at their
site. This would have to be scalable to handle lots of small
companies otherwise it won't turn a profit. The contractor would
then sell the service of an old PC running the monitoring software
and reporting to Novell's system every minute or so (a stable
Internet connection would be needed). Novell's software would,
automatically, call the cell phone of the contractor and one other
person (perhaps the business owner) when a problem was detected
(hard drive space, tape backup failed, etc).
The business owner is happy because his systems are being
monitored and the contractor is happy because he gets billable
hours and Novell is happy because they get $10/month/site for doing
almost nothing. Like I said, this isn't Microsoft-level profits,
but it gets Novell some market share and it gets them inside
businesses that they would not have gotten into otherwise. Roles
involved: vendor, consultant, exec.
How about a completely different example, but along the same
lines. A vendor should have and maintain a database of contractors
and their area of operations and expertise. This is useful for when
a potential customer calls. Novell has the platinum/gold/silver/etc
resellers in their database. Now, include a section of "charitable"
options (hours available per week/month and categories such as
"church" or "political" or whatever). A lot of charity work is tax
deductible. Why shouldn't the vendor be nice and match a non-profit
organization with a consultant looking for some deductible hours?
Again, this is about market share, not massive profits. But it does
get the vendor's name known and liked. Roles involved: vendor,
consultant, exec (of the non-profit).
A third example: slowing down the competition. There are still
lots of Exchange 5.5 servers out there, not to mention all those
small companies using MSN or AOL for "company" email. Why not help
them out while saving them the time/money of upgrading to the
latest version of Exchange?
- The vendor packages a "smart host" system and the contractor
deploys it to handle the spam/viruses/etc for the Exchange 5.5 box.
It should be all Free/Open software so there shouldn't be any
charge for the software.
- The vendor provides a smart host outsourcing service (#1,
except no new hardware/software at the client's site). This way,
the vendor gets a monthly payment.
- The same as #2, but the Exchange server's data (not the box or
the software) is absorbed by the vendor and the clients connect
Again, the idea is to provide functionality and stability to the
client for as low a cost as possible. Throw in as many extras as
possible, without increasing the cost. Then hand off as much of the
customization to the contractor as possible. Roles involved:
vendor, consultant, exec.
Right now, Linux is growing from the world-wide vendor selling
to the large companies and from the individual techs deploying
individual solutions for individual companies. Why not combine the
vendor's world-wide scale with the contractor's regional experience
to sell services to the small companies? It would require a
different form of partnership between the contractors and the
vendor, but both roles (and the client) would benefit from it.