Getting Aboard the Information Access Platform
Delivering the Capacity to Connect Users, Work Processes and Data
If you are the CEO of an IT consultancy gazing out over your customers’ application landscape, here is what you see: valuable information buried in piles of documents, databases, email messages and “stovepiped” applications. Even when relevant knowledge is explicitly documented, identifying it is becoming more and more difficult due to the growing flood of mostly irrelevant information. Highly paid knowledge workers spend much of their time looking for needed information, and essential know-how often exists only in the heads of a few employees. This is in stark contrast to the considerable efforts that have been made to improve production workers’ productivity by placing every tool and part they need within easy reach.
Obviously, a huge potential for improvement exists... and there is no lack of potential software tools and technologies. Enterprise content management systems (ECM) offer integrated suites that cover most or all of the content lifecycle from authoring/scanning to syndication and archiving. However, from an enterprisewide information access standpoint, there is at least one major problem: ECM systems are document-centric, but vital corporate records also include structured data in relational database tables and business applications (e.g. ERP, CRM).
Portal software brings together functionality from diverse internal systems, but this tends to expose merely the more visible elements and services of individual systems (“portlet profusion”)—rather than unifying content to allow access in the business context. Industry experts generally point users to search engines to obtain a more integrated view of heterogeneous repositories.
Enterprise search engine technology is undergoing a quiet evolution. Several years ago, the industry seemed obsessed with translating the value proposition of popular Web search engines into the enterprise. Today, search engine technology is becoming more diverse. The emerging trend is to move from standalone applications or appliance models toward an information access platform—technology components that you can assemble in flexible ways to bring content from multiple applications together in order to deliver information needed in critical work processes. This is a kind of next-generation middleware; a new software layer, or “enterprise bus,” that helps glue together heterogeneous information from both content stores and database-backed business applications.
Why An Information Access Platform?
There is an increasing realization that today’s document processing technologies are limited in that they hardly capture the wealth of knowledge contained in documents, email messages and database fields:
- For natural-language documents, often the information is not completely spelled out on the linguistic surface and must be retrieved by interpreting the text within a broader understanding of the world (“world knowledge”); and
- Capturing information from structured applications like databases, ERP or CRM systems is limited by the need to understand application semantics—information contained or implicit in a particular database field or application object.
To extract the information contained in documents or structured application data, we first have to identify the various kinds of knowledge that give rise to a particular document or database field, such as objects, actions and properties. The initial set of entities collected from the document corpus will have only restricted coverage—it contains only those concepts that actually occur in the documents themselves. There are obvious gaps where related concepts are missing, where an enumeration is not complete, where properties are not specified, etc. In order to arrive at a more complete picture of the work domain, one needs to take additional sources into account: experts, textbooks, product models from industry, etc. All of these have to be coordinated to form a coherent whole. At present this can only be done manually and it requires a measure of centralized management, workflow and expense that is beyond the grasp of many enterprises.
Fortunately, analyses of knowledge needed in business processes often reveal opportunities for specific, point-to-point gluing together of information where the interrelationships among data and processes can be more narrowly defined and therefore more easily tackled. Rather than centrally connecting many diverse repositories to a search appliance, or tearing down legacy applications and centralizing content in an all-encompassing ECM system, one can build separate search-based applications focused on specific business processes. An information access platform can reduce the amount of code that developers must write to create such search-based applications. It offers:
1. A search engine (sometimes called “search bus”) with open API architecture as the glue between the user and the data in diverse enterprise applications. It unifies the data by structuring it within the context of a specific business process;
2. Secure connectors that absorb the heterogeneity and volumes of your information systems;
3. Result presentation in a common Web style that is intuitive to all users; and
4. Monitoring, control and high availability/failure recovery.
The move away from standalone applications to a platform model is great news for IT consultancies, VARs/system integrators and software vendors. The platform model enables them to reduce the amount of code developers have to build for search-based applications. It provides a common meeting ground where IT consultants and developers can deliver value to enterprise users, with the architecture of the platform and the APIs forging the critical relationship. And it provides software vendors with a common business foundation that can serve multiple customers.
Oracle’s strategy is to build its Secure Enterprise Search product into a powerful platform for information access by leveraging our position in technology infrastructure—offering a complete chain encompassing the hardware, operating system, portal, content management system, database and business applications.