Network Visualization and Service Monitoring Research in a Large ISP Stephen North Information Visualization Research AT&T Laboratories, Florham Park, NJ (USA) north@research.att.com © 2008 AT&T. All rights reserved. Experience Vizgems Impact Graph visualization technology Version 1: 300,000,000 voice call attempts/day; 1 feed Improved readability of network diagrams Version 2: 500,000 virtual circuits; 5 OSS feeds Large network rendering and browsing Version 6 - 7: supports large enterprise customers “Graphs as maps” 50,000 Internet hosting devices Hybrid techniques for analysis + visualization 12,000 WiFi access points ICDN Vizgems - managed service visualization Enables data integration in the human interface. Shows service configuration and status in realtime. (planned) 50,000 managed security service routers AT&T Domestic IP backbone, 2007-8 • ~ 800 high speed routers Scales to hundreds of millions of events/day. • ~ OC768 (40Gbps), OC192 (10Gbps), OC48 (2.5Gbps) backbone link speeds Integrates archival and online data. • ~1500 GB/day in detailed measurement traffic (Netflow) • Common Backbone has ~ 18 TAP collectors, each monitoring ~40 backbone routers • Busiest collectors handle a daily average of 25 Gbps of traffic Provides point-and-click maps, charts, tables, and network diagrams in generic web pages. QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. © 2008 AT&T. All rights reserved. QuickTime™ and a decompressor are needed to see this picture. Network visualization myths Network visualization is about making new graphical metaphors. Network visualization is about making it easy to access data and see patterns. Network visualization should contribute powerful new tools. Visualization should contribute effective techniques to other platforms. Network management relies heavily on comprehensive overviews for awareness of system status and analysis of emerging problems. Network management is organized around specific, focused tasks. Network operators need tools for automatic anomaly detection and prediction. Network operators need tools for anomaly understanding. Databases and data warehouses make it possible to cope with scale in visualization. Well…. © 2008 AT&T. All rights reserved.