The acquired company, which numbers all of 16 employees, adds the ability to infer relationships between metadata. That is, if A relies on B and B relies on C, Celebra’s technologies concludes that A and C also have an interdependent relationship.
It works the same way with transforms, making it possible to draw more conclusions about which services could be reused for a new or newly orchestrated business process. And it could be used for gauging the impacts of a change in a particular service or business process.
Business analysts won’t have to navigate through complex hierarchies to reuse services, explained webMethods CTO Marc Breissinger.
webMethods was originally planning to embed the technology as part of an OEM relationship in the forthcoming webMethods 7 product, which is tentatively set for year-end release. Although webMethods never publicly announced the OEM plans, it did previously disclose them to customers under NDA.
The technology utilizes several W3C standards associated with the Semantic Web, including Web Ontology Language (OWL) and the Resource Description Framework (RDF). However, although they have been officially ratified, neither standard has entered wide use, given the underwhelming response to make the web itself more semantic.
Consequently, while use of these standards might provide a degree of interoperability with semantic frameworks that operate over the web, for now that remains a big if.
Instead, webMethods prefers to emphasize that this is not about the semantic web, but instead, making services networks themselves more semantic or meaningful. These standards are invoked under the covers, and users do not need to know them.
What’s interesting is that the technology recalls similar attempts to craft inference languages during the abortive emergence or artificial intelligence technologies 20 years ago. There were many problems with early AI, among them weak hardware and primitive techniques for so-called garbage collection, that eliminated irrelevant associations. But the biggest problem of all was that classic AI applications required complex, specialized languages because the rules or inference bases had to be built by hand.
The difference this go-round is that the Cerebra technology being acquired by webMethods is fully automated, and won’t require customers to learn specialized languages. Instead, it will operate using wizards that either automatically define semantic relationships, or provide the means for users to specify them.
Initially, webMethods plans to incorporate the technology into its SOA Fabric product, where it will be used for spotting opportunities for reuse ore provide grist for change impact reports. It could peruse listings residing in service descriptions in local or remote UDDI web services registries.
While not planned for the next release, webMethods also plans to extend the technology to discover relationships in applications linked through webMethods’ more traditional EAI adapters, such as in SAP systems.
The 16-person company, based in Carlsbad, California, was acquired for an undisclosed, modest sum, and the deal is already closed. WebMethods will continue to operate the offices as the place where it develops the semantic SOA technology.
