GNOSIS: Global Network Operations Status Information System

Monitoring the global state of a network is a continuing challenge for network operators and users. It has become still harder with increases in scale and heterogeneity. Monitoring requires status information for each node and to construct the global picture at a monitoring point. GNOSIS, the Global Network Operations Status Information System, achieves a global view by careful extraction and presentation of locally available node data. The GNOSIS model improves on the traditional polling model of monitoring schemes by 1.) collecting accurate data 2.) decreasing the granularity with which network applications can detect change in the network and 3.) displaying status information in near real-time. We define the Network Snapshot as the basic unit of information capture and display in GNOSIS. A Network Snapshot is a visualization of locally available state collected during a common time interval. A sequence of these Network Snapshots over time represent the evolution of network state. In this paper, we motivate the need for a network monitoring system that can detect global problems, in spite of both scale and heterogeneity. We present three design criteria, Accuracy, Continuity and Timeliness for a global monitoring system. Finally, we present the GNOSIS architecture and demonstrate how it better detects network problems which are currently of concern. The goal of GNOSIS is to present a stream of consistent, accurate local data in a timely manner. Comments University of Pennsylvania Department of Computer and Information Science Technical Report No. MSCIS-01-27. This technical report is available at ScholarlyCommons: http://repository.upenn.edu/cis_reports/155 GNOSIS: Global Network Operations Status Information System Jessica A. Kornblum and Jonathan M. Smith {jkornblu, jms)Cldsl. c i s .upenn. edu Distributed Systems Laboratory. Computer and Information Science Department University of Pennsylvania