Performance monitoring

These webpages are a place to find information about the status and the work of the performance monitoring group.


Introduction

Intra- and Inter-domain monitoring infrastructure aims to monitor "performance" metrics (such as one-way delay, jitter, packet loss, available bandwidth, etc) inside a domain and across several domains. The aim is to provide to the different groups of users (NOCs, GRIDs, etc) an "across domains" view. The measurements are collected and stored in a domain and exchanged when requested by another domain.
This activity will first focus on defining which subset of metrics should be monitored, the interactions between the domains and provide some suggestions on how the metric should be monitored (see
activities list). An embryo of inter-domain monitoring infrastructure should be setup by Q3 2003.

For this purpose, the group will build a "domain central tool" to retrieve domain measurements and exchange these collected measurements with other domains when requested. Recommendations on monitoring tools will also be made.
The architecture will be tested during a trial phase. During this phase, some measurement boxes and the "domain central tool" will be installed on at least three adjacent domains. The metrics monitored during the trial are the One-Way Delay (OWD), the Instantaneous Packet Delay Variation (IPDV), the One-Way Packet Loss (OWPL) and the packet re-ordering. The main focus will be on these metrics, but the architecture must be build in such a way that additional metrics can be added.

Later on, the group will focuse on the integration of the other metrics and the development of a tool to detect and diagnose where the problems are located. Those tools should be tailored for network manager (independently of where they are) and if possible to the end-users.


Documents

Architecture v0.1 doc pdf
Domain tool v0.1 doc pdf


Presentations


Performance Monitoring Meetings - Minutes and Presentations


Contact: Performance monitoring mailing list or Nicolas Simar