Clever Recordkeeping Metadata Project Outcome Area 5

Service orientation for recordkeeping

More flexible and dynamic recordkeeping in the Internet-enabled world is imperative in order to achieve interoperability, as highlighted by the opportunities for recordkeeping in true service oriented architectures. Significant exploration of this environment by recordkeepers is essential to enable appropriate responses. As an interim step the use of limited web services (rather than fully transformed service oriented approaches) to wrap legacy systems (hardwired solutions) will enable some immediate returns.

Brief discussion

The outcomes of building a limited functionality metadata broker in the middleware environment quickly demonstrated the requirements for greater flexibility, extensibility and robustness to generate dynamic metadata translations. The emergence of truly service oriented architectures (SOA) within the computing field, while generating much hype, and possibly 5-10 years away from full realisation, offers great potential to re-think and re-engineer recordkeeping functionality, processes and delivery in this environment. SOA offers the potential ability for recordkeeping processes to be orchestrated into business processes driven by recordkeeping requirements articulated in rules bases. In the interim, however, the availability of the CRKM Metadata Broker model as a web service offers immediate relevance to organisations seeking to improve recordkeeping interoperability between known systems.

Summary findings

  • Conceptualisation of records services in the SOA environment is at a much greater degree of granularity than required in the paper world – for example, portions of services will aggregate to form a records process (translation services are only one component of records capture, and similarly a component of a number of other records processes).
  • The conceptualisation of recordkeeping will affect how the articulation of services is defined and realised.  For example we hypothesise that definitions of records services such as ‘capture’ articulated by NARA in their FEA services for records management, embodying a life cycle concept, will be different to those articulated using a records continuum conceptual approach.
  • We anticipate significant functionality, yet to be articulated, associated with records services in particular areas. In particular, the use of rules based services may be particularly useful for the expression of mandate and recordkeeping events. Monitoring services also seem particularly likely. These are not a part of electronic recordkeeping functionality commonly specified today. However, a service component approach to defining monitoring services may involve areas such as checking formats (how many Word 2 docs left?) for migration processes, or monitoring changes to systems residing external to the records system used to source metadata values - for example, monitoring an X500 directory for any changes to position, contact details etc.
  • We hypothesise, and find emerging evidence to support the hypothesis, that defining recordkeeping as service components will alter the architectural options available to recordkeeping – both potentially away from monolithic records or content management applications, and also in providing radically different recordkeeping options to implement recordkeeping functionality through flexible interfaces to business application systems never designed to be records systems. The following diagrams illustrate recordkeeping as utility services within a service oriented architecture and how these recordkeeping services could be conceptualised from a service oriented perspective.

Recordkeeping service bus

Fig 5-1: Service oriented architecture for business processes

Recordkeeping data/information/object stores

Fig 5-2: Service oriented architecture for recordkeeping