An operating system tells a computer where to find data, and where to find various services, such as drivers for your screen. It points to them.

The Internet acts very much like an operating system, everything on the Internet points to something: pages, data, images, services such as search, maps, etc.

In the world of operating systems there are proprietary and open systems.

- Microsoft's Windows family is proprietary and Linux is open. You can make changes to Linux as long as you share the changes and keep everything open.

- Open operating systems provide a commons where everyone shares in improvements but it is more difficult to make money from open systems.

Hybrid OSes...

You can have hybrid operating systems such as Apple with its Mac OS X, which uses a version of Linux (BSD), combined with proprietary software. Apple saves money on core OS kernel development and it can focus on the proprietary modules where it builds value.

The Internet is very much like a hybrid operating system, making use of core, open standard technologies such as the Internet Protocol, HTTP, HTML, CSS, XML, etc. These are mixed with proprietary technologies such as Adobe's Flash, Google's Search, Facebook's Social Graph, MPEG-2 video, etc.

And the race is on for companies to try to own as much of the Internet operating system they can manage.

Tim O'Reilly, the head of tech book publisher and events organizer O'Reilly, has written a great essay that describes this competitive race. In State of the Internet Operating System Part Two he evaluates the competing contenders to the Internet Operating System throne.

Amazon, Apple, Facebook, Google, Microsoft, and VMware all have credible platforms with strong developer ecosystems. Then there is a collection of players with strong point solutions but no complete operating system offering.

Mr O'Reilly does a good job in analyzing each company, what it has and what's missing. No one company comes out with everything--at least not yet.

He writes that the alternative is to have the Internet become a collection of services connected with open standards.

I've long used Tolkien's "one ring to rule them all" as a metaphor for platforms that seek, like Windows before them, to take control of the entire developer ecosystem, to be a platform on which all applications exclusively depend, and which gives the platform developer power over them. But there is another alternative. Both Linux and the World Wide Web are examples of what I call "small pieces loosely joined" (after David Weinberger's book of the same name). That is, these platforms have a simple set of rules that allow applications to interoperate, enabling developers to build complex systems that work together without central control.

This is the way things work today: small and large pieces loosely joined. This is the Internet that startups want because it helps to create a level playing field with open access to data and services. But this is not the reality we see.

The trend is for companies to try to own as many "pieces" as possible, or to have as many companies as possible use their "pieces" through APIs (Application Programming Interfaces) -- such as Google with its Maps API.

The problem with building a business with APIs is that you are at the mercy of the API provider, they can share a lot of their service or restrict it with little warning. Ross Mayfield, from SocialText calls this strategy "Ajar"-- the APIs are partially open.

Mistrust...

While APIs are a great way to have companies use your "pieces" and develop new applications around your technology, there is growing mistrust of this strategy by startups. For example, Twitter recently become a competitor with many of its application developers using its API. And this same scenario will be repeated repeatedly by other companies.

This will affect how the "small pieces loosely joined" Internet develops, or not. This ideal Internet is disappearing and being replaced by a familiar pattern of consolidation and greater aggregation of value by each of the large Internet players. It's the same pattern we saw in the PC market.

Time will tell...

While no company is close to having all the parts of an Internet operating system there is another way to look at this scenario and that's to see who has the most attention of Internet users.

And Facebook gets a lot of user attention and in my opinion is the best placed to become the Internet operating system for many users.

Take a look at these numbers:

- Nielsen calculates that in March, 2010 US Internet users spent 1 hour 19 minutes on Google.

- Those same users spent 7 hours on Facebook.

Google is better at monetizing attention but Facebook will have plenty of opportunities to monetize all that user attention.

Facebook is busy creating the parts of the Internet operating system that it lacks, and it partners with companies for the rest. Eventually, it can fill in all the parts it needs. This is the company to watch.

In the meantime, all the other large Internet players will be trying to own as much of the Internet operating system that they can manage -- which means there will be lots of developers and companies that get caught in the middle. Some developers will do well by being bought, others will be left to struggle against huge competitors that can scale competing services across a massive user base.

Size matters...

Scale is rapidly becoming the differentiating factor that determines Internet success. While this has always been the case to some extent, as the industry consolidates and user bases grow to ever larger levels, the scale of the Internet players becomes huge and hugely important.

The difference between a competitor that has 20 million users and 50 million users is large but it's not insurmountable, Facebook has 400 million users -- that's a tremendous barrier to competition.

This next stage of the Internet will gradually become less startup friendly as the big players suck up more of the value creation.

Take a look at what happened in the PC industry. Intel and Microsoft managed to use their scale (and proprietary technologies) to suck up most of the value creation, they regularly report profit margins of 60 per cent plus... while PC makers and others in the PC ecosystem, survive on single digit profit margins.

That's the same trend we see developing in the Internet market. The race is on to aggregate the most value (profits) while using tremendous scale to protect against competitors.