Background

CAIDA's border mapping tool bdrmap tackles the problem of automatically and correctly inferring network boundaries in traceroute. bdrmap focuses on the network boundaries for which we have the most confidence that we can accurately infer them in the presence of sampling bias: interdomain links attached to the network launching the traceroute. Our method infers all interdomain links directly connected to and visible from the network hosting a single VP. Accurately inferring the parties involved in all interdomain links observed in traceroute requires overcoming the natural sample bias in traceroute, i.e., poorer visibility into distant networks, which limits our ability to assemble constraints. The bdrmap process builds on years of prior work in topology discovery, alias resolution, AS relationship inference, and active probing systems.

Methodology

The bdrmap approach begins with assembling routing and addressing data used to inform data collection and analysis. In particular, we use the following input datasets:

• Prefix-AS mappings constructed from public BGP data (routing table snapshots from RouteViews and RIPE RIS projects). We also used CAIDA's AS-rank algorithm to infer AS relationships for the same BGP data.
• Public datasets supplied by the five Regional Internet Registries (RIRs) that report address blocks they have delegated to networks.
• A list of IXP prefixes from database snapshots provided by the PeeringDB and Packet Clearing House (PCH) projects.
• A list of sibling ASes of the network hosting our measurement VP. We seeded our manual inference with CAIDA's public AS-to-organization mapping, and then manually added missing siblings and removed spurious siblings.

Then, we deploy an efficient variant of traceroute to trace the path from each VP to every routed prefix observed in the global BGP routing system. We apply alias resolution techniques to infer routers and point-to-point links used for interdomain interconnection. We use this collected data to assemble constraints that guide our execution of heuristics to infer router ownership. We developed a specialized measurement utility that we call bdrmap to drive data collection and infer border routers. The goal of bdrmap is to obtain as much information available about the links observed from a given network toward every other network, in order to constrain our subsequent border router inferences. We implemented bdrmap as a driver to scamper, a parallelized measurement system that efficiently gathers raw traceroute and alias resolution data.

The entire border mapping methodology including data collection and analysis phases is described in the paper, "bdrmap: Inference of Borders Between IP Networks, M. Luckie, A. Dhamdhere. B. Huffaker, D. Clark, K. Claffy, ACM SIGCOMM Internet Measurement Conference (IMC) 2016.

Data Format

For each VP, the dataset contains one file representing the output of the border mapping process performed from that VP during the month. The format of the file is:

<IP>[,<IP>]|<AS>:<IP>[,<IP>,]|<AS>:<IP>[,<IP>]

Each line of the file represents a border router inferred as owned by the network hosting the VP, and shows all routers adjacent to this border router. The fields on the line are separated by "|". The first field is a set of IPs representing a router owned by the network hosting the VP. Subsequent fields represent other routers connected to the border router owned by the VP network. Each subsequent field consists of the AS number of the neighbor network operating the attached router, and the set of IP addresses inferred to be on that router.

Acceptable Use Agreement

Access to these data is subject to the terms of the following CAIDA Acceptable Use Agreement

When referencing this data (as required by the AUA), please use:

The CAIDA UCSD Border Mapping (bdrmap) Dataset,
https://www.caida.org/catalog/datasets/bdrmap_dataset/

You are required to report your publications using this dataset to CAIDA.

Request Data Access

• Request Access to the CAIDA Interconnection Datasets.