Networks go peer-shaped

The way computer systems are built could be set to change with the arrival of the peer services process model.

Every now and then the way computer systems are built changes significantly. It moved from batch systems to online systems with the advent of minicomputers, and changed to client-server systems with the advent of PC networking.

This model has persisted on the internet, in the form of browser to web server, with a huge rise in connectivity.

Now this model for building systems may be about to change. That, at least, is what Tim Negris, chief executive of Equinom, is arguing.

Negris has produced and published what he calls a Peer Services Manifesto, which suggests that a new model for software development will emerge.

If you want to read the manifesto in full, it can be downloaded from Equinom's website. If you just want the headline ideas, then here they are.

The manifesto's basic premise is: "The client-server technology model, in all of its forms (including the internet) is inappropriate for the network business model."

The reasons it gives are as follows:

• Client server does not allow direct technology costs to be shared between the different entities participating in a business operation or transaction.
• It forces the arbitrary aggregation of private information into a common database whose operator must be trusted to protect and manage it, and not to misuse it.
• It cannot take advantage of the ever-growing processing power and storage capacity of desktop computers.

The author uses the term 'network business model' to refer to business activities that work through the loose association of multiple entities: people or small companies, or even large ones.

Interestingly, these are the least automated business processes because they are inter-company rather than intra-company. They are supply chain transactions of various kinds.

The manifesto maintains: "The network business model needs a different technology model for effective automation, which includes: equal or proportionate direct technology cost sharing among participants; owner-controlled information visibility, security, privacy and use; and distributed workload sharing across participating systems."

The author observes that technology vendors currently price, sell and value their products in a way that makes sense only to the world's large corporations, and claims that this acts as a brake on the automation of network transactions.

He then describes how emerging technologies such as web services, grid computing and peer-to-peer (P2P) capabilities are providing the foundation for a peer services architecture.

At first it seems that Negris and his company are simply jumping aboard the latest bandwagons and trying to hijack them. Until, that is, the manifesto proposes a 'peer services process model' and a 'peer services data model'.

These are distinctly new ideas, to me at least. The peer services process model envisages distinct peer processes on a peer grid, registering what they can do with the grid's operator and then seeking complementary services to carry out business processes.

Peer processes can thus find and consume services and also offer services themselves.

This may sound odd until you realise that it is the way in which open outcry markets, for example, work except that it is people rather than software that make it happen.

The peer services data model envisages data being primarily private, unless it is required to be used collectively.

This is distinctly unlike the database as we know it, although it does accurately reflect the data privacy interests of individuals or companies in a shared environment.

Naturally, Equinom has an axe to grind. It was formed to build and realise a peer services architecture and it intends to bring some form of development software to market to capitalise on the direction it believes the IT world is bound to take.

Until we see the technology, it is going to be difficult to know how this vision stands up. Equinom is promising to give its software away to a few carefully selected organisations, such as trade associations, schools and possibly law enforcement agencies.

It claims that many such organisations have opportunities to build useful applications using peer services, and the company intends to seed the market.

But Equinom is not an open source operation and, for most organisations, the technology will have a cost.

Equinom argues, however, that the applications built using its technology are destined to be inexpensive both to deploy and to run.

After all, they will be running mostly on the configured desktop computers across broadband connections, and the cost of any application will be shared among its users.

For me the emergence of a respectable P2P software movement is a bit of a surprise. When Napster first appeared and glowed like a shooting star, there was a great deal of enthusiasm for the idea of P2P systems.

But nobody stepped up to provide the P2P idea with technical respectability. This has now changed.