Next Generation Information Server: Traits Explained

The Next Generation Information Server is being built around the concept of data model composition through so called Traits.  This video explains the base principles of Traits.

About these ads

About jeroenvr

CTO & Distinguished Engineer EMC Information Intelligence Group
This entry was posted in Crazy Content Vlog and tagged , , , , , . Bookmark the permalink.

8 Responses to Next Generation Information Server: Traits Explained

  1. Hello Jeroen.

    I would like to know if a trait could be added to an object types definition or just to a specific instance: could you explain more, please? Is the “trait” concept similar to a TBO or to an Aspect?

    I published a link to your video/post on my personal site ( : this page (and I suppose the video too) has been viewed 30 times since yesterday; this number is about half of the viewers reported by YouTube (67 viewers). I am happy to have added visibility to your post.

    Cheers from Italy,
    Yuri Simione

    • jeroenvr says:


      Thanks for posting my video. I appreciate it.
      In NGIS each object is composed out of Traits and there is no other way to model metadata. Traits are similar to Aspects but they are a first class citizen and have more configuration and enhancement capabilities. Aspect in Documentum for instance have a weaving mechanism behind the scene to determine ordering of event handling. In NGIS we make this configurable.
      Also the metadata model in NGIS is much more powerful compared to Aspects. Schema (XSD) driven metadata definitions for Traits is really powerful.
      Traits are attached to Objects at runtime. We expect that there is a need for optional traits , mandatory traits and system traits where you could define at the object definition level which traits are mandatory to auto populate traits when objects are instantiated.

      Hope this helps.

  2. Sandor Bonnema says:

    Hi Jeroen,

    Good to see that the datamodel is going to be schema based and modular. This will simplify the use of XPlore for storing and searching metadata.
    I have one question: when I add an attribute to a schema, will there be support for updating existing object instances as well or will they lack by default the new attribute?



    • jeroenvr says:

      This is an interesting question, thanks for the response. I should write an entire blog entry on upgrades of data models since it is quite neat.
      In the Next Generation Information Server we allow you to modify your trait schema and define so called upgrade or migration policies. A migration policy associated with a trait definition tells the server how to handle metadata model changes of the existing instances with traits associated with this schema. We intent to implement various upgrade policies such as: upgrade existing objects with new default values, upgrade existing objects with some calculated value, don’t upgrade existing objects etc.

      We specifically want to support multiple versions of types running side by side allowing you to NOT upgrade objects. This doesn’t apply only to the metadata model. It also applies to the associated services.

      This can lead to all sort of interesting combinations such as; backwards compatible services enhancements over multiple versions of the data model aka upgrade the service but don;t upgrade the metadata, not backwards compatible service enhancements where the old version of the service is running against objects with the old version of the data model, not backwards compatible service enhancements where the metadata model is upgraded before the new version of the service runs against it.

      We also demonstrated at EMC world that you can set policies that determine when upgrades occur. In a cloud platform it is extremely useful to do rolling upgrades of your metadata model while using the system. We are designing the Next Gen Information Server to truly work in a 24×7 environment where developers do frequent upgrades of the system and its metadata data without any down time.
      Hope this helps.


  3. Pingback: Things and numbers from the EMC IIG Q4 2011 earnings release | Information Organization & Access Today

  4. Pingback: Traits in #NGIS: Cautious Optimism on the #Documentum Back End | kominetz

  5. Nic says:

    Will Traits be a replacement for, or an addition to the various BOF constructs – TBOs, SBOs and Aspects? Traits look very powerful, but many customers have large historic investments in current Documentum technologies – CS and BOF (also WDK, but I’m less concerned about that).
    If Documentum’s existing customers are forced into a large scale redevelopment to migrate from CS to NGIS, it’s inevitable that they will be obliged to undertake a formal assessment of NGIS in comparison with other competitive offerings available at that time.

    • jeroenvr says:

      Thanks for the question. We really see a co-existence of the current content server and NGIS with the ability to gradually move to a multi-tenant architecture. We realize that the investments in the current content server are significant and we want to innovate at the same time. In my upcoming EMC World sessions I will spend time explaining this strategy. If you cannot make it then you can always listen to the recordings. So Traits are a different and more powerful approach to metadata modeling and behavior over metadata.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s