Riding the “Enterprise Information Bus”

[Latest News Headlines]

27th October 2009 heralded the start of the EA Roundtable series in the UK with the first discussion focused on Managing the Information Explosion Challenge. The customer group comprised of Enterprise Architects from a diverse range of industries ranging from Transportation and Public Sector to Financial Services  assembled at  Bafta’s prestigious Headquarters in London.

There are many ways in which this Information Architecture focused topic could be interpreted (see related blog on this subject by Hamza Jahangir) ranging from the management of ever increasing data volumes (e.g. a key factor underpinning many new industry initiatives such as Smart Metering or simply the proliferation of business/social collaboration) to an integration/search perspective where careful consideration is required around how you bring all related forms of information together (structured/statistical or unstructured/semantic) to make it more useful and meaningful to the information consumer. Our discussion focused more on the latter than the former but also in the context of:

• The changing nature of business operating models and growing decentralization of data sources
• The challenges associated with governing unstructured data sources whilst still trying to apply the same architectural principles as you would expect with structured data
• The criticality of unstructured data to core business process within an organization

Whilst it was generally accepted that there was an equal explosion of structured and unstructured data, the debate focused on the fact that businesses seldom make best use of unstructured data outside of the context of major business applications (e.g. ERP or CRM). This is largely due to unstructured data being more subject to interpretation and that without some form of role-based semantic filtration, its true value/potential may never be realized. Moreover, the semantic nature of unstructured data makes it more difficult to govern compared to its structured counterpart thus questioning its intrinsic value and provenance – put simply, can businesses trust the data ?

A mash up is typically associated with the fast/simplified integration of information from multiple data sources. However, they are often hand-crafted in terms of binding structured and unstructured data sources for a specific purpose or information consumer role rather than being dynamically generated based on a query based (ad-hoc) interpretation. To that extent, not all information (stored in a diverse range of formats – e.g. Video, Blogs, structured database tables) may be included within a given query even though all this information could be semantically linked. Are we therefore truly harvesting the value of the Information Explosion ?

Despite all of these issues, the group agreed that it would be useful for MIS systems to query/search both structured and unstructured data sources to enrich the value of corporate MI.



The concept of the “Enterprise Information Bus” was discussed as a data services driven mechanism (enabled via Service Oriented Architectures) for provisioning both traditional statistical data sources (e.g. packaged application schemas or data warehouses) as well as semantic data structures . The complexity of binding the data sources would be abstracted from the information consumer who would merely see the bus as a dynamic information provisioning device. Furthermore data could be dynamically pushed onto the bus as well as pulled but the logic of semantic association would again be abstracted from the information provider. Finally, the idea of a '”Semantic Repository” was also put forward as a method of adding semantic structure to unstructured data sources that would help the Enterprise Information Bus to associate the information with structured data. Cool stuff eh ?!

Finally, the group also debated the issue of data sources being distributed across multiple infrastructures (i.e. some on customer premise, others hosted within public or hybrid clouds) as a by-product of the Information Explosion and the knock on impact  of being able to efficiently collate/provision this data whilst maintaining the business and architectural principles of integrity/trust and security etc… which is a great lead into our next EA Roundtable topic on Cloud Computing on 18th November 2009 …



More from the Enterprise Architecture at Oracle Blog ...