What Oracle multi-tenancy means for PLM providers?

Oleg, nice views. I like to think of a data problem from three angles, and personally don’t think there is a silver bullet. Ideal solution would be a combination of three IMHO. Multi tenancy is one design, which I would personally think is more important from a SaaS/PaaS provider’s point of view. For an organization having a private cloud for it’s DB multi tenancy is rather lower priority. In this case the other design comes in play- sharding. Can you effectively split your data across multiple shards (machines) so that you can process separately and thus scale by adding more nodes in your farm. This will require some non-relational and non-linear thinking

Third component/design – which I am not aware of in depth but roughly thinking of- ability to query on comparitively large amount of unstructured data. I don’t know if this should be Lucene or Hadoop or something else. Thoughts based on my experiences, open for correction of course!

Oleg, I am sure you are aware that Oracle purchased Endeca Enterprise Search technology some time ago. Similar to Microsoft sorting out how to apply Fast to their ECM strategy (with Sharepoint 2013 they seem to have finally nailed this), I believe Oracle are now working on how to use Endeca’s toolbox across the enterprise (ERP/ECM) and also for a future cloud strategy play. Endeca was always very strong in terms of offering both structured and non-structured search, correlation and ‘on-the-fly’ facet/drill-down navigation. Personally I find it curious that Autodesk, Dassault and PTC are apparently incorporating search engines within their PLM applications while Oracle and Microsoft seem to be on a path to become dominant enterprise search platform vendors in this area. Surely from a corporate perspective, CIOs will push for PLM to publish to Enterprise Search / ECM platforms instead of requiring PLM consumers in the organization to ‘occupy’ a PLM license to consult and search PLM data? I can see challenges in presenting BOM data and 3D models within a corporate Enterprise Search context, but perhaps an alternative path will naturally arise in developing ‘containers’ for this sort of data to be ‘searchable’ within this sort of context? Thoughts?

Oleg,

Sorry for the confusion, but your post was not clear in terms of what info you were requesting regarding Oracle’s non-structured data approach, hence my reply, but on a side note Enterprise Search will also have to adapt to a multi-tenant architechture in the future…perhaps this could lead to something interesting in terms of the trade-off between ‘sharing’ relevance search, dictionaries, taxonomies between companies in a particular industry segment on the cloud X data privacy.

BTW I was fully aware that you had founded Inforbix, and I have been tracking your technology since 2011. I even discussed in a meeting in São Paulo, Brazil with Jim Quanci (Jan 2012) that Inforbix could be an aquisition target for a major PLM player, such as Autodesk.
I still believe that within the corporate environment the enterprise search platforms (Microsoft/Oracle) will dominate as a ‘top’ search layer . Perhaps PLM search will be one more repository for a federated corporate search crawler and feed into this corporate layer. In other words offering a richer user experience within the PLM tool, while at the same time preparing the data for exposure in the corporate environment.
Then again, you might also be looking outside of the corporate environment at the future crowd/cloud scenario where a cloud based PLM search engine would be very interesting for thousands of small companies / inovation startups that need to digital protoype, speed up design re-use and time-to-market.
Cheers,
Bruno