APART-2 workshop on Grid Monitoring and
Performance Analysis

Session APART.1: Performance Analysis for Grids
Session APART.2: Performance Monitoring for Grids
Round Table Discussions

Session APART.1: Performance Analysis for Grids

Time Tuesday, August 26, 2003, 09.00-11.10
Location E 2-42
Chair Emilio Luque
   
09.00 - 09.40 MPI performance prediction by DIMEMAS
    Rosa M. Badia (Barcelona) invited speaker
09.40 - 10.10 Automatic online performance analysis for Grids
    Michael Gerndt (Munich)
10.10 - 10.40 Towards a Grid performance analysis concept
    Nemeth Zsolt (Budapest)
10.40 - 11.10 Grid Benchmarking
    Pedro Trancoso (Ciprus)

Top

Session APART.2: Performance Monitoring for Grids

Time Tuesday, August 27, 2003, 11.30 - 13.10
Location E 2-42
Chair Michael Gerndt
   
11.30 - 11.50 Grid resource and application monitoring by Mercury
    Gabor Gombas (Budapest)
11.50 - 12.10 Combining R-GMA, GRM and Mercury to provide a generic Grid monitoring infrastructure
    Norbert Podhorszki (Budapest)
12.10 - 12.30 Application Monitoring and Performance Analysis in CrossGrid
    Roland Wismueller (Munich)
12.30 - 12.50 Grid and java monitoring
    Marian Bubak (Krakow)
12.50 - 13.10 Towards a Unified Monitoring and Performance Analysis System for the Grid
    Linh Truong (Vienna)

Top

Round Table Discussions

Round Table Discussion 1

Time Tuesday, August 27, 2003, 14.30 - 15.50
Location E 2-42
Chair Peter Kacsuk
Thematic:

Relationship of Grid resource monitoring, Grid application monitoring and Grid performance analysis

Questions:

1. Do we need separate Grid infrastructure monitoring and Grid application monitoring? If yes, what are the main distinguishing features of them?

2. What is the meaning of performance in the Grid? At which level should we define Grid performance: application, network, site, VO, etc.?

3. What is the meaning of application performance in the Grid? How can we mmeasure it? Can the Grid programmer influence the application performance in the Grid or is it the infrastructure that shoud take care of the mapplication performance in the Grid.

4. What is the real Grid users' view on Grid monitoring? What is their experience?

5. Is there any major missing funcionality in the current Grid infrastructures that prevent generic Grid monitoring and/or Grid application monitoring?

6. Is it enough to use only the Grid information system (like MDS-2 or R-GMA) for Grid monitoring or do we need other Grid components for Grid monitoring and/or Grid application monitoring?

Round Table Discussion 2
Time Tuesday, August 27, 2003, 16.10 - 17.30
Location E 2-42
Chair Marian Bubak
Thematic:

Relationship of Grid resource monitoring, Grid application monitoring and Grid performance analysis

Questions:

1. Why are there so many parallel, redundant, duplicated and uncoordinated R&D activities in our field? How could we achieve a better coordination among the exisiting Grid projects?

2. How could we put together the results of Grid monitoring achieved so far into a usable Grid monitoring infrastructure?

3. How could we create a registry of existing Grid monitoring projects, Grid monitoring products and Grid monitoring research results?

Top

Back to Program Overview