Building Nemo, a system to monitor IP routing and traffic paths in real time

Efficiently operating and quickly troubleshooting Internet backbones require timely and accurate information about the routing system, their traffic demands, and how they are routed. Unfortunately, the dynamism and complexity of routing protocols, their lack of monitoring features and intricate inte...

Full description

Saved in:
Bibliographic Details
Published in:Computer networks (Amsterdam, Netherlands : 1999) Netherlands : 1999), 2016-03, Vol.97, p.1-30
Main Author: Raspall, Frederic
Format: Article
Language:English
Subjects:
BGP
Computer networks
Internet Protocol
IP (Internet Protocol)
Legacy systems
Management
Mathematical analysis
Mathematical models
Monitoring
Monitoring systems
Networks
Protocol (computers)
Real time
Routing
Routing (telecommunications)
SNMP
Spatial distribution
Studies
Traffic engineering
Traffic flow
Traffic information
Trajectories
Troubleshooting
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Efficiently operating and quickly troubleshooting Internet backbones require timely and accurate information about the routing system, their traffic demands, and how they are routed. Unfortunately, the dynamism and complexity of routing protocols, their lack of monitoring features and intricate interaction, together with the volume of routing state make their monitoring network-wide (e.g. across an AS or ISP) hard; while the narrow scope of traffic data provided by existent measurement tools and scalability issues make it difficult to infer the spatial distribution of the traffic. We present the design of a Network Monitoring tool (Nemo) conceived to offer two critical functions: first, monitoring nodes’ routing and forwarding state in nearly real time; second, the ability to derive forwarding paths, which can permit obtaining spatial representations of the traffic when combined with traffic measurement data. We discuss the hardest challenges to overcome to support these functions, how we addressed them, and show preliminary results obtained in two operational networks to illustrate the feasibility and utility of the tool. Nemo was conceived to work with legacy equipment, relying only on standard, widely-supported functions. Nevertheless, we identify downsides of recent monitoring features and suggest simple modifications to router behavior that could greatly ease monitoring the routing system.
ISSN:1389-1286
1872-7069
DOI:10.1016/j.comnet.2015.12.011