Changes between Version 17 and Version 18 of GEC19Agenda/DeveloperTopics


Ignore:
Timestamp:
03/25/14 11:20:10 (9 years ago)
Author:
tmitchel@bbn.com
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • GEC19Agenda/DeveloperTopics

    v17 v18  
    5252 7. TSM; Tools to Scale up Experiments (Marshall Brinn, GENI Project Office)
    5353 8. ABAC Support; Status, Plans and Benefits (Marshall Brinn, GENI Project Office)
     54
     55== Session Summary ==
     56 1. Brecht Vermeulen, UGent and Fed4Fire
     57  - Brecht discussed and demonstrated the tools available in the [http://jfed.iminds.be jFed toolkit]
     58  - Brecht described the progress on the Uniform Federation API and the Common Aggregate Manager API
     59  - Brecht provided lessons learned from the first operational cycle of Fed4Fire
     60 2. Slice Exchange Point and Common API v2.0 - Michiaki Hayashi, KDDI
     61{{{
     62#!comment
     63  - Michiaki gave an overview of the Slice Exchange Point and Common API v2.0, including a description of the architecture and walk through the documentation
     64}}}
     65  - Michiaki discussed his recent work on Slice Exchange Point (SEP) and the Common API v2.0. The SEP has played a key role in building global slices and experimental topologies, making disparate resources from distributed heterogeneous testbeds behave as a single  integrated slice. He describe a set of resource descriptions per testbed that allows for this kind of "infrastructure neutrality". He reviewed the progress he and his team have exhibited from a concept presentation at GEC14 to demonstrations at GEC17/GEC18 of a VNODE/PG slice. He reviewed the newly documented Common API v2.0: Federation Architecture and Common API / Common Slice Definition that supports and extends the kind of interoperability he and his team have been exploring. We discussed the similarities between his Common API and the GENI AM API and how we might be able to bridge between these in the future.
     66 3. Update on Federation API - Tom Mitchell, GENI Project Office
     67  - Tom spoke about the benefits to tool developers of the Uniform Federation API
     68 4. The GENI Aggregate Manager API; Ratify and Discuss Changes (Aaron Helsinger, GENI Project Office)
     69  - Aaron described a series of GENI AM API changesets that had been under discussion for some time and asked that they be ratified. The assembled audience briefly discussed and ratified the following for inclusion in AM API v4 and optional inclusion in APIv3:
     70   - [wiki:GAPI_AM_API_DRAFT#ChangeSetN:AddinformationtoGetVersion Add information to GetVersion]
     71   - [wiki:GAPI_AM_API_DRAFT#ChangeSetO:Refinecharacterrestrictions Refine character restrictions]
     72   - [wiki:GAPI_AM_API_DRAFT#ChangeSetP:SupportproxyclientsthatSpeakForanexperimenter Support for speaks-for clients]
     73   - [wiki:GAPI_AM_API_DRAFT#ChangeSetQ:Supportchangingusersandkeysonexistingcomputeslivers Changing users and keys on existing compute slivers]
     74  - Aaron provided an overview of other changesets for possible discussion during the coding sprint (time permitting):
     75   - [wiki:GAPI_AM_API_DRAFT#ChangeSetT:LongLivedSlices Long Lived Slices]: This discussion was continued at the coding sprint.
     76   - [wiki:GAPI_AM_API_DRAFT#ChangeSetS:MoreinfofromGetVersion AM URN in GetVersion]: The group decided that an aggregate indeed has a single URN, which for some aggregates may be the same as the single `component_manager_id` from the Ad RSpec.
     77   - [wiki:GAPI_AM_API_DRAFT#ChangeSetV:CreateandManageDiskImages Managing Disk Images]
     78   - [wiki:GAPI_AM_API_DRAFT#ChangeSetAB:CaseSensitivity Case Sensitivity]: The group agreed that tools and aggregates should not change, so slice and project names in GENI will be case INsensitive, though services should aim to preserve the requested case.
     79   - [wiki:GAPI_AM_API_DRAFT#Deferred:ChangeSetO3:Allowunicodevalues Allow Unicode]: The group agreed that we do not require this functionality in GENI today. Revisit this in the future if the need arises.
     80 5. Stitching is Here; How Can We Improve It? (Aaron Helsinger, GENI Project Office)
     81  - Aaron discussed several ways to improve stitching now that it is in use by more experimenters, including:
     82   - Standard Error Codes from aggregates would help stitcher communicate common errors to the experimenter
     83   - Support VLAN tag negotiation and AM API v3, to allow fast agreement on free VLANs
     84   - Better Tool Support is needed like graphical tools, possibly supported by a new service or library to encapsulate the workflow logic
     85   - AL2S will replace ION and help address many current issues, we just don't know when
     86   - Stitching all nodes at an Aggregate to the same inter-AM VLAN requires aggregate support, but would enable more complex topologies
     87   - Implementing Stitching Schema v2 will allow us to implement !OpenFlow controlled VLANs and later possibly multipoint VLANs
     88  - We continued this discussion at the [wiki:GEC19Agenda/CodingSprintTutoring#SessionSummary Coding Sprint].
     89 6. Speaks For: Status of Roll Out and Plans (Tom Mitchell, GENI Project Office)
     90  - Tom gave an overview of speaks-for authentication and proposed a roll out plan:
     91   - GEC 19: infrastructure (control frameworks, testing tools, common codebases)
     92   - GEC 20: aggregates
     93   - GEC 21: tools
     94 7. TSM; Tools to Scale up Experiments (Marshall Brinn, GENI Project Office)
     95  - The session ran out of time but Marshall discussed this topic at the [wiki:GEC19Agenda/CodingSprintTutoring developers coding sprint]
     96 8. ABAC Support; Status, Plans and Benefits (Marshall Brinn, GENI Project Office)
     97  - The session ran out of time but Marshall's slides are attached