Workshop Series Program Outline
Overview
First Workshop ( March 16-18, 2004 ), SLAC
- Application-science requirements and opportunities
- Technology/CS status and futures
- Gap analysis
Draft report on these topics to be produced by April 9.
Extended Organizing Committee Meeting ( April 20-22, 2004 ), SLAC
- Develop straw-man strategies for meeting needs
- Assume no, or very weak resource constraints, but do estimate resource needs.
- Describe strategies in terms of program solicitations aimed at any appropriate combination of CS, Industry and Office of Science Applications.
Draft report to be produced by May 14
Second Workshop (May 24-26, 2004), Chicago
- Present and refine or revolutionize strategies for meeting needs.
- Get feedback from the science applications about the likely relevance and effectiveness of each strategy.
- Develop at least a prioritzed overall program
Final report to be produced by August 15.
Detailed Outlines
First Workshop
Assume 18 schedulable hours (in 3 days)
Focus on:
- Technology status and futures
- Application requirements and opportunities
- Synthesis of requirements
- Gap analysis with respect to technology futures
Time breakdown:
- Welcome, Introduction etc. 1 hour
- Application Needs (6 hours):
- Office of Science applications. Two panel discussions, probably separated by the non-SC presentations below. Brief presentations based on list of questions. Important to understand workflow of each application. Panel chair to steer the discussion to identify similarities and differences.
- ASCI, DOD, NASA-EOS, .... Presentations focusing on the same list of questions.
- Technology status and futures (6 hours):
- Tape storage
- MAID (massive array of idle disk)
- Disk futures
- Memory futures
- Overview of file systems Lustre etc.
- Future of database technology in data-intensive science
- Network – opportunities and needs relevant to data management
- Grid – opportunities and needs relevant to data management
- Information management technology
The presentations should aim at giving science applications an understanding of what is likely to exist.
- Synthesis and Gap Analysis by technology capability (4 hours, probably parallel, no more than three parallel streams, hence combine some of the topics below)
- Archival Storage (no access, no power)
- Powered-on storage
- File Systems
- Database Systems
- Networks
- Grids
- Information Management systems
Gap analysis should focus on both performance and function
Summaries (1 hour)
Extended Organizing Committee Meeting - SLAC
Assume 16 schedulable hours
The meeting will be face-to-face and will involve the Organizing Committee members plus any application, CS or industrial participant they decide to invite. This meeting is expected to involve far fewer people than the two workshops.
Focus on designing gap-filling programs of work involving some or all of science applications, CS and industry. The programs should be described in terms of draft program solicitations, including funding. There is only a weak imperative to keep the total cost realistic and multiple approaches to the solution of a problem are encouraged.
Time breakdown:
- Introduction and recap of the first workshop gap analysis (2 hours)
- Decomposition of problems into technical breakout groups (1 hour)
Candidate areas (to be mapped on to ~3 groups):- Hardware oriented
- Low-level software oriented (e.g. file systems)
- Database development/enhancement
- Data transport and caching, networks and Grids
- Information management
- Breakout groups
- Analysis of technical and CS issues (2 hours)
- First pass at program solicitation (2 hours) If a particular area does not merit a program solicitation the reasoning should be recorded.
- Presentation and discussion of first-pass program solicitations (3 hours)
- Breakout groups
- Create draft program solicitations (3 hours)
- Presentation and discussion of draft program solicitations (2 hours)
- Wrap-up and marching orders (1 hour)
Second Workshop - Chicago
Assume 16 schedulable hours
Focus on taking the Gap Analysis from the First Workshop, and the ideas from the Extended Organizing Committee meeting, and producing plans that have the consensus support of science applications and computer science. Plans should be formulated in the form of program solicitations identifying the program elements that are designed to meet critical needs and the elements that are designed to address opportunities that must not be missed.
Time breakdown
- Welcome, introduction, recap of first workshop focusing on gap analysis (1 hour)
- Presentation and discussion of each draft program (including arguments for minimal or "leave-it-to-others" programs . (3 or 4 x 1.5 hours)
- Discussions to be panel discussions involving relevant application scientists who have prepared initial responses to the draft. Example program topics (to be combined into no more than four sessions):
- Programs focused on commercial developments
- File system technology
- Database Technology
- Data transport and caching Interface to Grid technologies
- Metadata and provenance
- Parallel breakout sessions: (3 hours)
- Refine or revolutionize draft programs, paying particular attention to realistic costing. Arrive at the prioritization of programs from the point of view of each application science.
- Presentation of the results of the breakout sessions (1.5 hours)
- Discussion of possible overall picture (with discussion leader) (1 hour)
- Parallel breakout sessions: (2 hours)
- Firm up the prioritization and costing of scenarios that meet the overall needs of science applications.
- Summary presentations and Wrap-up (1.5 hours)