The macroscopic behavior of the TCP congestion avoidance algorithm
ACM SIGCOMM Computer Communication Review
Using pathchar to estimate Internet link characteristics
SIGMETRICS '99 Proceedings of the 1999 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
An investigation of geographic mapping techniques for internet hosts
Proceedings of the 2001 conference on Applications, technologies, architectures, and protocols for computer communications
Topology discovery for large ethernet networks
Proceedings of the 2001 conference on Applications, technologies, architectures, and protocols for computer communications
TOPOMON: A Monitoring Tool for Grid Network Topology
ICCS '02 Proceedings of the International Conference on Computational Science-Part II
Multivariate resource performance forecasting in the network weather service
Proceedings of the 2002 ACM/IEEE conference on Supercomputing
Forecasting network performance to support dynamic scheduling using the network weather service
HPDC '97 Proceedings of the 6th IEEE International Symposium on High Performance Distributed Computing
Direct Queries for Discovering Network Resource Properties in a Distributed Environment
HPDC '99 Proceedings of the 8th IEEE International Symposium on High Performance Distributed Computing
Synchronizing Network Probes to avoid Measurement Intrusiveness with the Network Weather Service
HPDC '00 Proceedings of the 9th IEEE International Symposium on High Performance Distributed Computing
The Architecture of the Remos System
HPDC '01 Proceedings of the 10th IEEE International Symposium on High Performance Distributed Computing
Network Characterization Service (NCS)
HPDC '01 Proceedings of the 10th IEEE International Symposium on High Performance Distributed Computing
Mesh generation and optimistic computation on the grid
Performance analysis and grid computing
Nettimer: a tool for measuring bottleneck link, bandwidth
USITS'01 Proceedings of the 3rd conference on USENIX Symposium on Internet Technologies and Systems - Volume 3
SPAND: shared passive network performance discovery
USITS'97 Proceedings of the USENIX Symposium on Internet Technologies and Systems on USENIX Symposium on Internet Technologies and Systems
Comparing Passive Network Monitoring of Grid Application Traffic with Active Probes
GRID '03 Proceedings of the 4th International Workshop on Grid Computing
Synthesizing Realistic Computational Grids
Proceedings of the 2003 ACM/IEEE conference on Supercomputing
Enhancing QoS metrics estimation in multiclass networks
Proceedings of the 2007 ACM symposium on Applied computing
Self-adjustment of resource allocation for grid applications
Computer Networks: The International Journal of Computer and Telecommunications Networking
SLA-based complementary approach for network intrusion detection
Computer Communications
Hi-index | 0.00 |
Because the network provides the wires that connect a grid, understanding the performance provided by a network is crucial to achieving satisfactory performance from many grid applications. Monitoring the network to predict its performance for applications is an effective solution, but the costs and scalability challenges of actively injecting measurement traffic, as well as the information access and accuracy challenges of using passively collected measurements, complicate the problem of developing a monitoring solution for a global grid. This paper is a preliminary report on the Wren project, which is focused on developing scalable solutions for network performance monitoring. By combining active and passive monitoring techniques, Wren is able to reduce the need for invasive measurements of the network without sacrificing measurement accuracy on either the WAN or LAN levels. Specifically, we present topology-based steering, which dramatically reduces the number of measurements taken for a system by using passively acquired topology and utilization to select the bottleneck links that require active bandwidth probing. Furthermore, by using passive measurements while an application is running and active measurements when none is running, we preserve our ability to offer accurate, timely predictions of network performance, while eliminating additional invasive measurements.