The (In)Completeness of the Observed Internet AS-level Structure

Abstract

Abstract-Despite significant efforts to obtain an accurate picture of the Internet's connectivity structure at the level of individual autonomous systems (ASes), much has remained unknown in terms of the quality of the inferred AS maps that have been widely used by the research community. In this paper we assess the quality of the inferred Internet maps through case studies of a sample set of ASes. These case studies allow us to establish the ground truth of connectivity between this set of ASes and their directly connected neighbors. A direct comparison between the ground truth and inferred topology maps yield insights into questions such as which parts of the actual topology are adequately captured by the inferred maps, which parts are missing and why, and what is the percentage of missing links in these parts. This information is critical in assessing, for each class of real-world networking problems, whether the use of currently inferred AS maps or proposed AS topology models is, or is not, appropriate. More importantly, our newly gained insights also point to new directions towards building realistic and economically viable Internet topology maps.

H-Pastry: An inter-domain topology aware overlay for the support of name-resolution services in the future Internet

“…The identification of peering links on the other hand, remains an open issue. Data traces produced by [25] and [26] diverge on the size of identified peering links, while the study in [27] argues that the methodology followed fails to reveal up to 90% of existing peering links.…”

“…Though this assumption adheres to the organizational or the administrative structure of several types of networks (e.g., university, corporate networks), it does not reflect more complex structures found in the Internet, such as multihoming and peering relationships (see Section II-C). Reports from CAIDA [25], research efforts in Internet cartography [26], [27] and large measurement studies such as [18] indicate that the Internet topology is evolving into a mesh graph dominated by peering relationships. Moreover, multihoming is a well established practice for traffic load balancing and reliability.…”

Cloud Computing for Global Name-Resolution in Information-centric Networks

“…Multihoming and the establishment of multiple peering relationships between ASes has resulted in an AS-level topology graph that is far from purely hierarchical, making the classification of ASes across the topology difficult. In this paper, we adopt the methodology proposed in [15] and also used in [14], according to which the set of ASes is classified into four tiers based on the number of their downstream customers (this number is called the customer cone or simply the cone of an AS). The four tiers are: (i) Stub networks, i.e.…”

“…It is clear that the vast majority of ASes belong to the Stub category. B. Data-Oriented Name Architecture 2 We omit relationships between ASes which belong to the same organization, as these have been reported to be very rare [14]. 3 We do not show the CDF for the entire range of cone size values (up to 35753 ASes) for visibility reasons.…”

“…As we saw in Figure 3, several ASes residing at lower levels of the hierarchy have significant resource requirements for the support of DONA, due to the existence of peering links with higher level ASes. Furthermore, according to [14], a significant fraction of the established peering links is not present in public BGP data, especially in lower tier ASes, where this fraction reaches up to 90% for large content providers. This means that the cloud will be critical in enabling DONA even at lower hierarchy levels.…”