Meeting Notes - 30.09.2011

Concerning assigned actions, it would be helpful to see the gaps within main areas: resource management (meta level), information, accounting, messaging, security, and federated access

  • Jedrzej Rybicki
    • starting from the "golden vision" we will derive some high-level objectives (e.g. unified security model, meta scheduling, etc).
    • this high-level objectives can be then transformed in some technical objectives (as we know them in EMI), this process is out of scope for the cloud task force, It is the job of EMI project management (technical director, strategic director, etc).

  • Shahbaz Memon
    • The current recommendation in the slides is lacking the vision and process of how we arrived at the conclusion of supporting service interoperation and appliance based use case.
    • In this meeting we agreed to revise the process of justifying our conclusions and directions. The current problem domain only focuses StratusLab project (which is a high priority as it is part of the collaboration roadmap + EMI management), but we still need to further analyze the missing pieces of other initiatives such as NIKHEF, CERN, SARA, and WNodeS. The problem domain is actually a center of gravity for our objectives. These objectives will bring the implementation scenarios to our EMI functional areas (compute, data, infrastructure, and security).
    • EMI is expected to enable scientific communities accessing multiple virtual infrastructure deployments. Although this is very ambitious goal to fit EMI in different technologies, the priorities will be to target implementations in the interest of Project.
    • Action: Gap analysis for StratusLab?

  • Owen Synge
    • notes and summarized "golden vision"?
      • Discussion of the PP summary slides. Much discussion was focused on why Shahbaz had focused on Stratus Lab implementation, it was established that although the group had wanted to remain abstract Shahbaz had been requested at short notice to be more concrete in his summary slides.
      • Implementation specific concerns. The group felt that focusing on an implementation was counter productive as the group believes that much like the Grid, the Cloud will be a heterogeneous collection of services. For sustainability reasons the group also felt that it was dangerous to tie work to a single implementation as, implementations would change considerably over time and the groups use cases that lead to specifications of work should not be effected by these changes.
      • Satisfying the objective of concrete recommendations. The Group decided that the correct way to address the criticisms of that the summary slides did not lead to a specification was to focus on a "Shared Golden Vision" and build a hierarchy of dependencies, and based upon this hierarchy look at the missing components, some of which may be beyond our current effort.
      • The "Shared Golden Vision" It seemed that all the group could agree on a "Shared Golden Vision".
        • Users should not only have to interact with a single Cloud, but could submit requests for any cloud that matches their requirements to a central service.
        • Users may instruct this service that they only want to interact with a single cloud.
        • That a user from a user community such as a VO should be able to submit a VM management request to this single service, and if this request was not tied to a specific cloud provider that the requested VM start request could be started by a provider that supported the user community or VO.
      • The beneficial side effects of the "Shared Golden Vision" It was seen by the group that the "Shared Golden Vision" brought together the usage scenarios presented by EGI and EMI management as key objectives, but made the issues more "concrete" such as clarifying what the federated Clouds where for with respect to the user.
      • Dependencies of the "Shared Golden Vision" It is seen that security, information systems and accountancy messages need to be standardised to reach the "Shared Golden Vision". Owen Suggested that we should consider some thing similar to the LCG Pilot Factory project, as a unifying basis for VM submission, and potentially a mechanism for validation of all concerns with federation and that after consulting with Markus Schulz was redirected to the maintainer Andrey Kiryanov
      • Realisation of the "Golden Vision" The group realises that the main purpose of the "Golden Vision" is for reference upon which Gap analysis, and functionality can be assessed. The smaller objectives of EGI and EMI can then be clearly understood even if completing the "Golden Vision" is after gap analysis discovered to be beyond what can be achieved in the lifetime of EMI.

  • Eric Yen
    • To provide gap analysis - ASGC cloud initiatives. (experiences from the deployment)?
    • Action: Will compile all the contributions to the document.

Conclusion

Revised after comments from others..
Edit | Attach | Watch | Print version | History: r9 < r8 < r7 < r6 < r5 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r9 - 2011-10-04 - MMemonExCern
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    EMI All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback