SON: Motivations

These are the motivations behind the Service Oriented Network.

Bad Business Models

Freeloading

A certain amount of freeloading will be tolerated on a network, so long as there is no return on stopping. In other words, if there is $100 worth of freeloading, but the lowest possible effort to stop it would cost $110, it's not worth the effort. Also, if there's $100 worth of freeloading and it would cost $10 to stop 5%, it also isn't worth it. This model is important because it was one of the first. The internet (and later the web) existed for academic research, but individual side projects were able to freeload because there was no point in stopping them.

Advertising

The basic model was context sensitive advertising. This took into account the content of the page you were on or the content of a specific query to deliver advertising.

This model was replaced with user-modeling where information is collected about a specific user so that advertising can be targeted to them. There's a part of this I don't like, but I'm not sure I can actually claim it's a "problem", at least not in the technical sense. Advertising is very effective. It exploits cognitive biases and social norms. Saying that someone "doesn't have to buy" is like saying someone "doesn't have to fall for a con".

There are two problems I can identify. The first is motivation or "market pressure". Any service earning it's profits through advertising will do better by increasing it's advertising capacity than by improving user experience (for the same cost). Over time, this erodes the value of the service to users in favor of value to advertisers. The second problem is that this model is the root of the NSA spying. These companies need to profile you for advertising, but those same profiles are valuable to intelligence agencies for spying. This makes it more economical for the NSA to spy on everyone than anyone.

Value Added Storage

Probably the least objectionable of the bad models. A service is provided (often on something of a trial or limited basis) and for a fee, you can buy some quantity of storage. Again, market pressures are now working against the consumer at some level. If a competing service arises, your current provide has a disincentive to make it easy for you to get your stuff out. This also messes with the value of storage. I haven't actually done the research (I should) but a very casual observation leads me to believe that the cap on the 'free' version is somewhere around 1/2 of what most people use. And the paid version offers something like 10x what most people use. I'd also guess that on tiered services, this rule holds pretty close for the division between tiers. The issue with this is that you may get 1G "free" when what you really need is 2G. But what you are sold is 10G. However, (and this is the sleight of hand), the price for 10G is really good, probably right on par with cloud storage provides ($0.10/mo AWS). But that's the wrong calculation, the right calculation is the value of the 2G you're actually using.

Finally, it reduces competition. The higher the switching cost, the higher the barrier to competition.

Digital Feudalism

Probably equally dangerous to advertising, and far more infuriating to me. These are the "app stores". Again, it kills competition because the switching cost becomes monumental. It's a game only the big players can get into. It's all around a great way to quash true creativity, and the creative destruction that comes with it.

Less Bad Business Models

Selling Software

People don't want to buy software any more. Those that do, pirate it. And business are motivated to produce incremental improvements, particularly when there is little or no competition.

Subscriptions

Ok when not associated with VAS. But still have a "handshake" problem. How do you convince the user your service is worth the price tag?

Solutions

As advocated in the SON, I believe the four components of applications (schema, service, logic and UI) should be separated so that each component has to compete independently. All software on the network should be open source. All infrastructure should be commodity.