110 likes | 236 Views
Group A (session 2). Q1. 1. How would we develop a (minimal) high level blueprint/framework/conceptual architecture or set of processes (or is there a better word?) to organize and coordinate the development and support of cyberinfrastructure , e.g. could expect that this would include
E N D
Q1 1. How would we develop a (minimal) high level blueprint/framework/conceptual architecture or set of processes (or is there a better word?) to organize and coordinate the development and support of cyberinfrastructure, e.g. could expect thatthiswouldinclude - minimal securityassurances, identitymanagement - datasharingpolicies - collaborativesoftwaredevelopment - campus bridging to international infrastructures - governancemechanisms
Q1 Discussion Notes (1) • Adoption and not re-inventing the wheel! • What good practices / processes have not yet been adopted by this community, but should be? • Not develop, but adopt! • Tried ITIL, does not really work for an evolving / dynamic environment, not easy to adopt • There are things which can be learnt from it though • Key players and stakeholders • What key players in setting that process up are required? • OSG, PRACE, EGI, XSEDE, ... (usualsuspects, but only e-Infrastructure/compute) • Whataboutusercommunitiesordataprojects? • How to identify others? Hierarchical approach for the process (too many)? • Potentially (too) many folks…to get consensus.
Q1 Discussion Notes (2) • Influence on governance and mission and the well-known legacy challenge • Communities have their own governance/mission, which will not change. • Communities will participate if there is an obvious benefit (and not before) • E.g. Does XSEDE know about community requirements, e.g. NCAR. yes, is documented. • Also have to deal with legacy... • Collaboration, risks and terms • Different funding sources seem to impose different requirements • different funding streams don't need to be competetive -> federation ofresources... • Risks for communities, e.gif NCAR participates in XSEDE, is it subsumed? • no, will maintain its identity – It needs to be well communities mutual benefits • Clear termsrequired: Research Infrastructure versus e-Infrastructure, tomake explicit which infrastructure is supposed to *serve* and those that are communities • Very different for HEP, because it is only one community specific CI, where value of overall project is more valuable than idendity
Q1 Discussion Notes (3) • Roadmapswithconcretetimelineandfollow-ups • On federations, policies etc. cancollide…in different projects • International / global roadmap for other (non-HEP) communities may help • That has to be (also) done independent of funding agencies (possibly some funding for doing so) • Culture of coming together is not bound to funding agencies, nor to funding, really • Example: humanities do not participate (lessfunding still) • Go beyond chasing funds, reconsider the fundamental requirements • Example: what are the key aspects of a federated identity system which we want? • *Then* discuss solutions...
Q1 Discussion Notes (4) • ‘Parked’ Funding for process to understand needs • Why do we do solutions before understanding the problem? • ECis not funding process, but proposed solutions with clear objectives (and KPIs). At time of proposal, it is expected to propose (targeted) solutions. ‘Parking money’ is usually not really appreciated • EarthCube, conceptualproposal • Open process though means x00 participants, which can get unwieldy • multiple funding sources further complicate that process... • This will converge if the value proposition is clear, and one can demonstratereturnofinvestment • What do you want to do -> what do you want to accomplish • Communicating vallue proposition can be difficult, takes time, needs patience • A wholepreparatory project orseriesofworkshopsmightbeinteresting
Q2 implied in Q1, time restriction 2. How would these processes aid in activities such as - sustainability - international cooperation - anyothers?
Q3 3. How to to turn this "blue print" into a set of actionable processes?
Q3 Discussion Notes (1) Approaches • 0: Within each community, community building beyond individual groups / institutions (similar to HEP, NCAR, others ) – ESFRI-like in EU • 1: need to agree that blueprint is needed. • 2: who owns process of creating the blueprint? • 3: how do we get the process of realizing the blueprint?
Q3 Discussion Notes (2) Is a blueprintneeded? • Blueprint for what? -> for global CI • Global may not be acceptable for communities, but rather federated, etc. • Sensitivity to language and terms, so define 'blueprint' etc before going to community • Establishglossaryofterms, communicateopenly • Small group, define process • Funding Agencies are stake holders, but they should not own the process (but helphereandtheretosupportthoseacivitieswouldbegood) • E.g. "A Blueprint for CyberInfrastructure for Global Science„ • That would, to some extent, be similar to defining the scope of EGI • Want members of community, not those who get at the end funded for doingit • separation of vision/design needs to be separated from realization/implementation • --> 1: yes, agreementthatthisisneeded
Q3 Discussion Notes (3) Who owns the process? • Small group, with input from others. • Somebody has to care enough about the bigpicture... • Blessing from Funding Agencies can help • How to fund this, e.g. PIRE??: Partnership for International Relationships • Orocess may work different in Japan, but may be adaptable • Needs vs. Desires must beunderstoodandderivedrequirementsare different too • Heterogeneity vs. simplicity? • Identifying needs takes time, lots of work • HUGE amount of work which is proposed here - Is itworthwhile? • Yes! • Then is matter of prioritization, human engineering • Requires significant time / commitment from key players (only those have the knowhow to get this achieved while at the same time being overcommitted) • You do not just hire new folks doing this