Architecture for Bring Your Own Service.

  • People are bringing more that just a device but people are showing up with their own services.  Regardless of what you bring, it will be integrated into the things we provide.
  • We are currently ready to adopt different services.  We only take on the things we bring in or bring up.
  • Case Study:** In eLearning you bring your own document store, portfolio and grade book and the LMS recognizes this and puts your files and outcomes into your personal stores too.
  • There are two contexts:  one is the thing you want to share and the other is?  
  • An open consumable API that would be published and anyone could plug into the API.
  • Is it taking the context of semantic web and turning into a semantic service.
  • You could take a your research processes and plug it into another University.   
  • Issues of trust when you bring services in and grant access to your personal services.  You also have trust of the systems of your stuff
  • There is both a push pull - a student might ask Siri what are the courses I should register for.
  • We could have a mash-up of services based on the API rules.   
  • What would be the principle first set of attributes you would need to have in the API?  
  • It must be fast and adaptable.   How can you be aware of and responsive of the ever changing device technology
  • There would need to be standards for the handshake.  Things like "I am a HIPPA Compliant Data Store".
  • This also applies to being able to pull down services and functions from our environment:  I only want the Approve Reqs out of PeopleSoft
  • This is the deconstruction of all the capabilities into little apps that can be mashed-up by the end-user.  The could mix their own capabilities with the universities.
  • Classification Schemes, Granularity, Taxonomy (HIPPA Compliant)

     Action Items

  • Develop an interesting article / maybe for EDUCAUSE Review
  • Review of who else is working on it.
  • Literature Review

How to use EA to Balance the Resource Allocation between Academic / Administrative (Marina)

  • Is this an activity across a PMO?  Do you parcel this out across projects via Project Allocation?
  • Academics - new requests for metadata archives, data archives and DBAs, app development, and the commodity 
  • There is a governance issue. A strategy issue. A resource issue because the academics are suddenly requesting resources they didn't request before.
  • How do we balance across the portfolios and leverage what we have already built?
  • You need to have a portfolio manage that spans the whole organization.  Right now it is done by each silo.
  • Marina should also add Internal IT to the Academic and Administrative.
  • You should make sure the governance people see the cost of the Run part of the resource allocation.
  • Enterprise Portfolio Management is all of these various slices - if you can populate the tool.   (Troux.com as an example tool).
  • You can deconstruct projects into capabilities.  If you build this capability first, you can leverage the capability across multiple projects.
  • Problem with PMO is they don't think of building blocks, they think of projects.
  • Sounds like they have an intake problem.   At UMUC they have relationship manager who is embedded with the unit and determines what level of work is (ticket or project).   Relationship Manager does all the ROI et al.  EA looks and says, we can do 80% of this with what is available.
  • Intake issues are on the Academic side. 
  • EA could frame the decision making framework and help fill out the Business Value Matrix

     Action Items

  • Come up with the principles and strategies that would be useful and guide the effort.
  • Come up with the tools that architecture can bring to the discussion.
  • See if there are case studies.

Capability In A Box (Jerome)

  • Can we roll-out modular capabilities at the System level that each campus can use either as a cloud-capability or pull down to their institution and run
  • Capabilities might be IDP, ESB, etc…
  • Rolling out a local independent capability at is not just a software thing but also a resource / staff thing
  • Could we do the 80% of the work at the core and leave only 20% at the edge.
  • Seems like an enormous savings opportunity.
  • Example:  Gluu will host and/or manage a shibboleth IDP for you.
  • Typically when you provide a service, you will have predefined integration points and data definitions.  That is part of the 80% of the work that would be done.
  • RedHat and IBM have "fabric" type offerings that do some of this.
  • This is a lot like the Bring Your Own Service - one is design and the other is deploy.
  • Leo - there was a California effort to do this many years ago.
  • In Germany, they created an Identity Management Appliance.  Many Universities came together to define and develop an IDM solution.   They built a package appliance that they took to the next institution.  The first university they plugged it into, took one year.  They documented how they did it.  The next institution took 3 months.
  • State California Initiative - SIF
      
       Action Items
  • Develop a set of criteria of what is a good candidate for a capability in a box
  • Match with the Bring Your Own Service and approach from the other side
  • Write a paper

Items we didn't get to:

  1. Rich's There are ideas for a Learning Activity Repository.   It would capture learning outcomes in the form: Noun (the learning), Verb (Mastered), Object (Accounting).  ITANA could work on the framework for the Verbs, and Objects.  
  2. Luc's Review of the landscape for Metadata automation and management for Research Data Archives.  
  3. Colleen's What are the professional development paths, goals and opportunities for raising a tech up to an architect.  What are the competencies you need to train up and skill sets.
  4. Leo's flesh out the IAM architecture for RATL model.
  5. Arin's how do you sell EA down to the on-the-ground activities?  How do you get them buy-into the idea of EA?
  • No labels