SOx in the City, Peter Golden

Peter presented an excellent overview of Web Services, part of the more general area of Service Orientation.

Peter’s slides were originally presented to a BCS Financial Services group meeting in the City with Laurence Julien of the Noel Group (http://www.noelgroup.co.uk). The slides are available from http://www.bcsfinsig.org/events/2003-04/031125.htm.

We learnt that Web Services are essentially a toolkit of technical infrastructure based on Internet standards, to enable computer-based services to be advertised, discovered and invoked. We were introduced to Web Services ‘alphabet soup’, and would-be WS bluffers will be particularly interested in some key acronyms:

• SOA (Service-Oriented Architecture): the architectural framework for web services
• XML (eXtended Markup Language): the language for defining any data associated with the SOA
• SOAP (Service-Oriented Access Protocol): the protocol to enable an SOA service to be invoked
• WSDL (Web Services Description Language): the language for describing web services
• UDDI (Universal Description, Discovery and Integration): a world-wide registry for recording and retrieving descriptions of web services.

Sufficient technology is already in place to make some use of web services. However, information security and gaps in the tool set are significant concerns, so realistically web services are currently being deployed mainly within corporate networks, closed to the outside world . Firewalls will need to become SOAP-aware before inter-organisation services can take place. This could put a brake on one of web services’ claimed major benefits, that of being able to invoke web services ‘on the fly’ from a free market of available services: it looks like users will need to talk to their security manager first.

The protocol stack misses some important communication elements, e.g. a reliable transfer service – although technology vendors are busy re-branding existing products to fill the gap, e.g. IBM’s MQ Series. Service repositories are expected to be in place in 2006. It’s believed that tools are presently good for support of web services clients but less satisfactory for servers.

Web Services do not define the functionality of specific business services, and are concerned only about mechanisms for invoking services. Most discussion centred on business considerations:

• Is there any ontology/taxonomy of web services? What models exist for syndication of information – use of information by one party of another party’s data? It looks like it’s up to business users to develop the necessary models, hopefully based on analyses already put in place, such as EDI.
• There could be an interesting role for third-party brokers – somewhat akin to broking between the parties in the liberalised electricity supply industry. Such brokers would act as intermediaries between the parties doing business, and could thereby help ensure reliable message delivery, provide records of transactions, etc. There would be a price for such services. All in all, there was a need for trust between the parties.
• Whilst the nature and main parameters of web services can be advertised, how can the contractual commitment associated with service provision be communicated? This seems to be an unexplored area as yet, and looks like a business challenge to be overcome if a free market in services is to take off.
• There was controversy over BT’s URU identity verification service, which operates by collating an individual’s personal information from a range of sources. The construction of this service is assisted by Web Services, but given potential Data Protection Act restrictions, the status of the service was unclear.
• There was much discussion of how quickly and successfully Web Services will be deployed. When and how will the critical mass necessary for widespread adoption of the technology be reached? Are Web Services a ‘disruptive technology’, i.e. a technology that completely changes the technology and business landscape? Is this an end-game technology? Is this just a technology-lead initiative (given that the likes of Microsoft and IBM are actually cooperating on the work as well as sinking huge sums of money on it), and are there major business benefits?

All in all, the feeling was that web services are a useful enabling technology, and that work is needed to incorporate an overall business architecture.