Regular Attendees
- Andrew
- Bill
- Brad but we will switch off so no one has to do too much.
- Chris
- Dan
- Jonathan (Presumed at CNI)
- Danny
General
- Call In To: Free Conference Call HD - DuraCloud Line
Please note that we will be using the Free Conference Call HD line for this call. Information about calling into this line is available from the link above.
- - Indicates who will be taking minutes
The meeting will be in two parts:
First Part 11-1
Break
Second Part 3-5
Discussion Topics
Useful Definitions
CSF = Critcal Success Factor
Risk = Not a negative but just a potential barrier to progress which needs to be identified and mitigated
What forms a critical code mass?
CSF - We should have an integrated product by the end of 0.2 upon which we can add features incrementally.
Question: What components/services are in this mix?
- one or more client sources
- full hook shot through DuraCloud, Fedora, and back into DuraCloud
- SI frontend
- security throughout
Question: What are the set of features per component we can do in 0.2?
Synchronization Service Watcher
- Additions to existing code?
Object Creation Service
- Convert to Apache Camel
- Deploy in Spring
- Add Droid (without OSGI this time?)
Question: What is the development/demonstration environment?
CSF - Demonstration Stories
Question: How do we explain what is working and what is to be developed?
CSF - Partners
CSF - UI Work
- Both for support base functionality?
- As a way to provide requirements and human factors information?
- Is the concentration mostly on the Synchronization Setup functionality?
Question: How do we want to engage researchers in 0.2?
Question: How do we want to work with SI in 0.2?
Risks (and how to mitigate them)
SI - What is there are problems integrating with SI from an information perspective?
SI - What if there are problems integration with SI from an security perspective?
SI - What do we do if we need to alter SI code?
SI - Can we demonstrate SI in an integrated fashion in 0.2?
SI - Is base Islandora a fallback?
Minutes
- Dan intros: Meeting in two parts; notetaking to switch off. Plan to cover the four existing cards, then move to new planning.
- Review of open existing cards for DfR 0.1:
- DFR-78 will be completed this week.
- DFR-23 moves to DfR 0.2, and will be factored into smaller tasks.
- DFR-74 DURACLOUD-575 closed today; waiting on DFR-76 (which went to DfR-0.2, and may move under epic). Closing DFR-74.
- Need to write up release notes.
- Goals for this meeting, and for DfR 0.2:
- Complete end to end system that we can then add features to.
- Need to clearly establish what can be built in the two month iteration.
- Expect design, storyboards, tech selection, wireframes; get validated by researchers in the project.
- Do DuraCloud deliverables represent a challenge for scheduling of DfR deliverables?
- (Discussion; will most deliverables be from Peter; can we get the planning done for Peter while the DuraCloud release is pending?)
- For 0.2 goal is to get a basis; get all major components of the path in place, expose any risks, and be prepared to add features.
- Do we want to cover the administrator or curator views in 0.2?
- Implied as part of basic account structure and identity and group management
- Rough out reporting and notification (are these the same technology base)
- Security for 0.2:
- All components with UIs will support Shib (entering user/pw flows throw shib provider).
- All service to service / non-browser interactions continue to flow via basic auth - with an access key associated with user account, not user/pw.
- Incorporate same into DuraCloud and DuraCloud management console and Fedora interaction and the SI UI.
- (DuraCloud was moving in this direction anyway for InCommon).
- Note that fine grained modeling of auth and associated UIs is expected to occur in a later iteration; determined by when user designs available.
- Not dealing with Fedora level fine grained access until later (but mechanism to be tested)
- SI Islandora for 0.2:
- Show basic tree of content.
- Not all SI content models supported. Possible stretch goal to have one data set that goes to SI content models and shows visualization.
- Can we show an exhibit display (Hydra/blacklight based) as a strong demo. (Can we get someone from Hydra to build this)
- We will need to choose the example data set and work towards the Fedora model to make that work.
- Search capability for a future version?
- Is this by 0.4, or 0.5 and beyond.
- Two-way or multiway sync by 0.4?
- NB difference between multiway sync and restore.
- (Discussion)
- Three Q's:
- Will we do multiway ever?
- How would we push augmented metadata back upstream (if we do)?
- What do we guarantee / not guarantee if the user aggressively creates an unsupported multiway environment?
- Decision: Place these into a phase two (not in DfR 0.1 through 0.4).
- Object characterization (e.g. format) in 0.2.
- Expect to add content models incrementally.
- 0.2: something; (one content model and examine Islandora to understand the risks). Also advanced processing in format characterization, ideally from a real data set.
- 0.3: add more, ...
- What do we need to do about async vs sync around the OCS etc for 0.2?
- Nothing? Small delays aren't expected to be a problem in the short term.
- Hookshot for 0.2:
- Researcher places content on local data system; watched by sync
- Sync pushes to duracloud
- Duracloud tells OCS, OCS pushes into Fedora
- Fedora sends object back via cloudsync.
- Can then find the file in Duracloud, see it in Fedora, and discover in SI frontend.
Need to define encryption strategy in 0.2 (strategy, not impl)
-
- OCS / microservices
Notes for Iteration Planning 2012-04-09
- DFR-74 - DFR012 - DF02
- DFR-76 Under Epic?
- DFR Chat Last Week.
- Expectations for Peter and Danny. What are needs to get started.
- Full core code
- Design, Wireframes, Technology Selection, validation with Researchers. DuraCloud deliverables.
- What time is available? From Danny, from Everyone Else.
- Bill Clients - Windows-client, Mac-client, Linux-client. Andrew meant what Bill meant.
- Brad - Adminstrator not called out. Curator. Accounts, Identity and group managements. Reporting framework and debugging not to become a risk.
- Notification - TBD
- 0.2 Diagrams need to be started. -TBD
- Security - How far in 0.2?
- All UIs have Shib!
- Service-to-service interactions have basing authN, authZ.
- RESTful basic authN.
- Bill: Follows for inCommon work.
- Andrew: Basic Admin, AuthZ for course-grained. Later, than for toward post 0.2.
- Jonathon: Content tree in demo in 0.2
- Brad: Stretch goal - hook to SI content model
- Dan: Just not SI.
- Bill: Other sorts of presentation - Look at Hydra/Blacklight post 0.2, Get help from Hydra
- Brad: Find a good data set underlying issue, finding might be 0.2 risk mitigation
- Dan: Private search post 0.2 post 0.3
- Synch tool: Additional capabilities post 0.4. Two way, Multi-way synch happens in the future.
- Brad: What if user starts doing multi-way synch outside the environment. What happens.
- Why would anyone put our synch tool in their environment.
- Future: Micro-service Execution Environment
- Bracket: In first forty minutes
- Format: As an 0.2 format function.
- Jonathan: More heterogenous demonstration
- Spell out - Full Hook Shot
- 0.2 Research puts content on local file system.
- To OCS.
- Back to DuraCloud.
- GUI - Designed - Using primitive UI.
- Bill: Stategy for Installing Monitor and Synch tooling, technology approach.
- Encryption: After 0.2. Define what we want to do.
MEE: OCS is steps within
1 Comment
Danny Bernstein
Notes for 3rd Hour
Metadata Collected at Ingest
Configurable Service Infrastructure
Notification
Search