In early 2001, British Telecom (BT), announced a software deal with a web services software vendor called Bowstreet, who were recently acquired by IBM. The purchase was made to create a European wide portal for the publication and distribution of web services. Incredibly insightful but the initiative eventually faded. Not long after I wrote a paper for BT on the 'anatomy' of a what is basically a web services broker. It's based on some earlier work I did with Anna Pollock, when we devised what a web services provider could look like in travel, by first creating a more generic model based on this diagram.
It's key features are;
- A directory of web services, community maintained by applying Wiki or ODP style usage protocols
- A supply side (institutions and independent developers) and demand side (end users and applications) presence
- An ontology (an explicit representation of shared knowledge) to make it easy to search and discover web services
- Core services that enable a provider to publish and/or host a web service and consumers to aggregate, syndicate and pay for services, together with a range of trusted services (encryption, billing, reputation metrics, etc)
- Facilities to enable fine grained services (e.g. post code lookup, price a bond) to be consumed by systems and applications, and coarse grained services (e.g. calendar widget, search widget) to be consumed by end users (web site, blog)
- Payment and revenue share services for providers and consumers (pre-pay & post-pay) that enable transactions between parties that are previously unknown to each other
- Community services that enable conversations between parties that are previously unknown to each other
Some friends at BT have been telling me about the resurgence of interest in this business model, the best example of which is StrikeIron, evidenced by BT's collaboration with Microsoft & TopCoder. They've been running software development competitions for independent providers where winners are provided with access to BT & Microsoft's customer base. Competitions are managed on behalf of companies such as Google and AOL (not unlike the Innocentive model in the pharma space). Interestingly, 45% of the individuals who enter are from India and 96%-98% of the winners are from China or Eastern Europe! There's an excellent article at Business Week covering the BT/Microsoft collaboration.
So what next? Getting the community dimension right cannot be overstated and continues to be a weakness for BT, although less so now for Microsoft. It's latest offering BT Workspace is too little, too late and misses the point completely, which is to put customers in touch with each other (enable conversations between parties that are previously unknown to each other). Without this BT will be unable to create liquidity in their fledgling marketplace where they already face stiff competition from Sourceforge (who have announced the imminent delivery of a marketplace) and others with existing critical mass (e.g. Xmethods).
BT, if you really wanna be starting something you need to see yourself as a market maker in the web services publishing business, bring some of your own assets 'out to play' and understand how to operate an online network, or face being just a tin and wires platform player.