Break-out Group Two (41GV)
Participants (41JY)
- What works? (41H5)
- Philadelphia Symphony is a new member (41IV)
- We can see who registers and is using perfSONAR (41IX)
- Asia PacificNet, LHC (Atlas more than CMS (41IY)
- DREN, I2, ESnet, IRNC networks, starlite, max, Ames - pushing toward all gigapops in US (41IZ)
- CANARIE was on DICE - now quiet (41J0)
- interested in circuit modeling (41J1)
- after GLIF - more interest (41J2)
- provide information to look-up register (generated by an entity) (41J4)
- many questions that network managers can't anwswer - expose the innards in a structured way at a level that serves the whole without divulging specific information that a party is not willing to share (41J5)
- engineers need a better method to discovery who to talk to fast to get to the root of the problem - this information would have been shared if you had a working relationship among the body of engineers - (41J6)
- institution decides what information is public and what is private (41J7)
- perhaps aggregation of peering points only (41J8)
- value of data already collected in standardinzed format (41J9)
- how much is exposed is a function of the trust levels (41JA)
- much more effective use of time by all involved (41JB)
- analogous to DNS - different view for internal and externsl (41JC)
- How to know what's there, how to articulate what I need, naturally expands to sharing of tools (41JD)
- major goal to allow existing tools to expose their data to a larger community - not to build tools (41JE)
- create underlying federated infrastructure, distributed databased with a mechanism for discovery (41JF)
- look-up - exists; toplogy - what might be relevant to problem at hand (41JG)
- improve individual and join trouble-shooting efforts (41JH)
- What else to include? (41JI)
- Could we automate not just finding but overcoming the problem? (41JJ)
- along the lines of self-healing computers (41JK)
- providing gobs of data may not be the way (41JL)
- Advanced dx tools give you more details, this architecture allows new data for research purposes to be gathered to continue to bootstrap - and later automate - and remediate in a need-driven, iterative way (41JM)
- perfsonar - framework for multiple data sources to build the context needed (41JN)
- Security - Shibboleth (41JO)
- Do any federal agencies use Shibbolethc? (41JP)
- Lack of community - everyone develops their own - How can we get the communities together (41JQ)
- Ken Klingenstein - Shibb - federated trust model - originating with I2 -university credentials re-used for other settings - federated - NSF for fastlane (41JR)
- DoD - PKI is dominant, HPC is kerboros (for users) - not sure what network people use (41JS)
- went from one (via kerboros - to many) (41JT)
- Can Shibbeloth interoperate with other ID mgmt. systems? (41JU)
- This approach should be seen as extensible - i.e. throughput between sites (41JV)
- How to move beyond existing developers who have evolved this organically in an open setting - where harder questions (security) deferred? Who has been forced to answer this? (41K8)
- Large science communities going in different ways - ESG, LHC, Ligo - working in smaller communities (41K9)
- In some settings very detailed information already collected, be aware of the challenge of usefulness of the data (41KJ)
- Anticipate a subscription model as part of the architecture? (41KK)
