WP 2 High Performance Networks

Last Updated: 9 November 2004

 Final Meeting : Final EU Review

Provisioning and Operations

Performance Monitoring

Grid Interoperability

Dissemination and Exploitation

Project Management DataTAG Partners

WP2 High Performance Networks : Objectives
The goal of this work package is the deployment of advanced network services for Grid applications across multiple domains for the optimal utilization of network resources. Network services include advance reservation and differentiated packet treatment, which is a function of the application sourcing data, while optimal utilization requires the tuning of existing transport protocols, elements of traffic engineering and  the identification and test of new ones.

Task 2.1 High Performance Transport
The goal of this task is the demonstration and deployment of high performance transport applications. for efficient and reliable data exchange over high bandwidth-delay connections.

Task 2.2 End-to-end QoS
The goal of this task is the study of the problem of Quality of Service provisioning in inter-domain heterogeneous environments

Task 2.3 Advance Reservation
The goal of this task is the evaluation of different advance reservation approaches and their interoperability between GRID domain

Task 2.1 Planning Documents
Task 2.1 Technical Presentations
Task 2.2 Planning Documents
Task 2.2 Technical Presentations
Task 2.3 Planning Documents
Task 2.3 Technical Presentations
Scalable TCP is a simple change to the traditional TCP congestion control algorithm (RFC 2581) which dramatically improves TCP performance in high-speed WANs.

Linux Kernel 2.4.20  technical report describes the structure and organization of the networking code of Linux kernel 2.4.20. including  the main data structures, the sub-IP layer, the IP layer, and two transport layers: TCP and UDP. 

Net100 The Net100 project has kindly make their latest patch available to us. It implements Sally Floyd's HighSpeed TCP proposal
High Speed TCP is a variant proposed by Sally Floyd which makes a simple change to the traditional congestion control algorithm. Over high-speed WANs, it allows faster recovery after packet loss, thus greatly improving throughput

"Benchmarking QoS on Router Interfaces of Gigabit Speeds and Beyond" in Proc. 9th Open European Summer School and IFIP Workshop on Next Generation Networks (EUNICE 2003) Budapest-Balatonfüred, Hungary, September 2003
"How To" Guides  Deploying TCP and QoS on Gigabit networks
GridDT proposes an alternative algorithm for congestion control in TCP. It dramatically improves TCP performance in high-speed WANs and addresses the issue of RTT bias.

Cumulative Patch incorporating HSTCP, Scalable TCP and GridDT

Packet Dropper for Linux

Fast TCP As a part of the DataTAG collaboration with Caltech, the Fast TCP implementation has been tested between CERN and sites in the U.S.
Meeting Reports
03-02-2004 07-02-2003
26-11-2003 05-02-2003
13-11-2003 06-12-2002
25-09-2003 29-10-2002
02-07-2003 01-10-2002
05-06-2003 Earlier Dates
26-02-2003  

EU Review, March 2003

Final EU Review, March 2004
EU Deliverables
D2.1   PM12 - Task 2.1 and Task 2.2
D2.2 PM24 - Task 2.1
D2.3 PM24 - Task 2.2
D2.4 PM12 - Task 2.3
D2.5 PM24 - Task 2.3

PFLDnet 2003, CERN
PFLDnet 2004, Argonne
PFLDnet 2005, Lyon

  GNEW 2004, CERN

WP2 Contact: r.tasker@dl.ac.uk

DataTAG is a project sponsored by the European Commission - EU Grant IST-2001-32459