Intersection Characteristics of End-to-End Internet Paths and Trees
ICNP '05 Proceedings of the 13TH IEEE International Conference on Network Protocols
Mapping and visualizing the internet
ATEC '00 Proceedings of the annual conference on USENIX Annual Technical Conference
Inferring subnets in router-level topology collection studies
Proceedings of the 7th ACM SIGCOMM conference on Internet measurement
Detection, understanding, and prevention of traceroute measurement artifacts
Computer Networks: The International Journal of Computer and Telecommunications Networking
P4p: provider portal for applications
Proceedings of the ACM SIGCOMM 2008 conference on Data communication
Brief announcement: the theory of network tracing
Proceedings of the 28th ACM symposium on Principles of distributed computing
SSS '09 Proceedings of the 11th International Symposium on Stabilization, Safety, and Security of Distributed Systems
Network Topology Inference Based on End-to-End Measurements
IEEE Journal on Selected Areas in Communications
On the hardness of topology inference
ICDCN'11 Proceedings of the 12th international conference on Distributed computing and networking
Misleading stars: what cannot be measured in the Internet?
DISC'11 Proceedings of the 25th international conference on Distributed computing
| Hi-index | 0.00 |
Computing the topology of a network in the Internet is a problem that has attracted considerable research interest. The usual method is to employ Traceroute, which produces sequences of nodes that occur along the routes from one node (source) to another (destination). In every trace thus produced, a node occurs by either its unique identifier, or by the anonymous identifier "*". We have earlier proved that there exists no algorithm that can take a set of traces produced by running Traceroute on network N and compute one topology which is guaranteed to be the topology of N. This paper proves a much stronger result: no algorithm can produce a small set of topologies that is guaranteed to contain the topology of N, as the size of the solution set is exponentially large. This result holds even when every edge occurs in a trace, all the unique identifiers of all the nodes are known, and the number of nodes that are irregular (anonymous in some traces) is given. On the basis of this strong result, we suggest that efforts to exactly reconstruct network topology should focus on special cases where the solution set is small.