Porridge Host
From Zanecorpwiki
(Note: company doesn't yet exist, but this is the idea.)
Development Hosting that's Just Right
Contents |
Goals
A hosting platform/community geared towards freelance developers and small development shops, the idea is that existing hosting solutions picture the market landscape as linear. At one end is the know-nothing developer who by choice or disposition knows nothing about hosting at all. They're able and willing to fit themselves to a strict regime, accept updates or lack-thereof for underlying tech, and generally do everything they can to stay away from admin tasks.
At the other end of the linear conception is the integrated developer who sees the Apache config files and Postfix mailing system as just another component in their byzantine development stack. The breadth of what they consider their purvey extends to everything.
With the rapid improvement of virtual machine technologies, a third leg in the equation has sprung up: expertise. Slice Host and Linode--for example--are both founded on the idea of you know how to do IT, so here's an easy way to get a flexible box without a big capital outlay.
The challenge is that these companies are one trick ponies. They provide a box on the cheap, but there's a difference between the group of people that *can* administer and those that *want* to administer everything.
Hence, Porridge Hosting. Like the DIY virtual shops, we offer simple virtual machines, but we also offer managed services so that if you don't need to customize a service, you don't have to. It's possible to mix and match managed services and custom services while keeping both development, production, and business infrastructures integrated.
Target Market
There are two general markets which Porridge Hosting has in mind.
Agile Developers Working with Whole Systems
We believe that for developers doing new work and especially those working in an agile mindset, this ability to take over and hand-off control of common services is essential. This is especially true if one works outside the business logic layer. In other words, when one is building a whole system (as opposed to doing grunt work on some part), then it may indeed become necessary to worry about a mail server. In other words, the question is whether or not just focusing on business logic is good enough.
Elite Web Users
The second group (which probably has significant overlap with the first) is elite web users. These are individuals that not only understand and use the power of the web, but have the capability to define what that power is. It's one thing to sign up for WordPress and quite another to write your own blog software. These men and women want the features they want and the fact that those features might not yet exist isn't going to stop them.
How It's Done
Many are familiar with the time-quality-features design triangle for software. Well, we view hosting as having three legs defined by ease-control-expertise. In the existing landscape, hosting companies clump around three spots:
- Cattle Hosts - no control, no expertise, high ease: give up all infrastructure control, fitting your development to the hosting company's standards, accepting upgrades or lack thereof as the host sees fit; useful for those that don't know nothing or don't want or need to mess with anything but pure code
- Enterprise Hosts - high control, no expertise, high ease: i.e., outsourced full on IT; pay a company a huge amount of money to do whatever you want without needing to understand how what you want is done; great if you can afford it
- Virtual Metal Hosts - high control, high expertise, no ease: geek virtual hosts; i.e., here's a virtual box, have fun
The Problem
The problem with the existing solutions is that they are both narrow and rigid.[notes 1] This is fine for the bulk of the market. Most customer's fit into one of the three areas.
But there is a significant--not so much numerically, but in terms of global value--demographic that has high expertise, but doesn't necessarily want the cost of some control to be total control. In other words, where does the developer than needs to control their mail service precisely, but doesn't care about the web server go? The Cattle Hosts are too confining, the Enterprise Hosts are too expensive, and the Virtual Metal Hosts are too time consuming. The cost of liberating one service shouldn't be the assumption of all services.
Notes
- ↑ Personally, I believe they are narrow because each vendor has chosen to see the world in a linear fashion. The no control, no expertise, high ease sees themselves as linearly opposed to the high control, no expertise, high ease, meaning, they only see the control dimension. The high control, high expertise, no ease see themselves opposed to both the previous groups and focus on the expertise. For them, control is a product of expertise, not an independent variable. These views of the world have created the narrow clustering of providers we see in the market and defined the possibilities in a rather rigid fashion such that they miss the opportunities afforded in the larger space.
