GGF 5 and GGF 6 Richard HughesJones The

GGF 5 and GGF 6 Richard Hughes-Jones The University of Manchester Data. Grid WP 7 Oct 2002 R. Hughes-Jones Manchester

GGF 5 u Network Monitoring Working Group u One working slot: n n n Discussion of Network Measurements document draft Network Characteristics hierarchy Measurement Methodologies and Tools lists Tools document Revisions to milestones u One speakers slot: n Henk Uijterwaal: RIPE Test Traffic Measurements project n Warren Matthews: IEPM-BW measurements and plans for MAGGIE n Richard for Tiziana : Closeness metrics n Rich Wolski: What do Short TCP Bandwidth Probes Tell Us? Data. Grid WP 7 Oct 2002 R. Hughes-Jones Manchester

Network Measurements Working Group u Details: Hierarchy Document/Future Plans (60) Presentations on Recent Work (50) u Chairs: Bruce Lowekamp, College of William and Mary Brian Tierney, LBL Richard Hughes-Jones, University of Machester u Purpose n Hierarchy document: Reviewed Network Characteristics Hierarchy, discussion focused on issues of Qo. S, Available Bandwidth, and Topology. Upcoming documents were revised and milestones adjusted accordingly. n Presentations: 5 Presentations were given on current network measurement systems and issues. u Plans n Hierarchy document will be completed and reviewed at GGF 6, submission to GFSG is expected at that time. n Survey of tools, and classifications of existing tools to characteristics is beginning, preliminary results and document expected for GGF 6. Data. Grid WP 7 Oct 2002 R. Hughes-Jones Manchester

Network Research Group u 10 things net people might wish grid people knew. . u At the receiver record n n n n Congestion Control - Slow Start is this always necessary? no ! Routing – let the Network people do it ! Packet Sizes Go faster LANs have always pushed the MTU up Overlays and P 2 p Qo. S - Would be nice (contrariwise: see Congestion Control) Multicast Tier 1 routing works. Most ISPs run core native multicast Layer 2 Considerations - layer 2 NBMA nets - lots - a pain Light v. Heavyweight Protocols Header prediction, Packet templates make stacks simpler n Macroscopic Traffic and System Considerations Self similarity, so? traffic is self similar (i. e. arrivals are not i. i. d) - this doesn't actually matter much. u 10 things grid people wish networking people knew n Help needed from Grid people !!! Data. Grid WP 7 Oct 2002 R. Hughes-Jones Manchester

GGF 6 u Network Monitoring Working Group u Session 1: Network Measurements Characteristics u Tuesday, October 15, 10 AM, Gleacher Rm 400 n Discussion of the Characteristics Document (to be linked in by Oct 4) is the primary task for this session. n Participants are asked to read the document in advance and be ready to discuss the proposed terminology, described characteristics, and what needs to be added or changed to make this a useful document. We are hoping that this document can be submitted to the GFSG before GGF 7. n Discussion of the next projects: tools survey, and relationships document, link to DAMED WG for publishing data u Session 2: Network Measurements Hierarchy and u Thursday, October 17, 8 AM, Sheraton - Mississippi n Wrapup of change or discussions on Characteristics document. n Planning for tools survey n Planning for relationships document Data. Grid WP 7 Oct 2002 R. Hughes-Jones Manchester