O Peer, Where Art Thou? Uncovering Remote Peering Interconnections at IXPs

Abstract

Internet eXchange Points (IXPs) are Internet hubs that mainly provide the switching infrastructure to interconnect networks and exchange traffic. While the initial goal of IXPs was to bring together networks residing in the same city or country, and thus keep local traffic local, this model is gradually shifting. Many networks connect to IXPs without having physical presence at their switching infrastructure. This practice, called Remote Peering, is changing the Internet topology and economy, and has become the subject of a contentious debate within the network operators' community. However, despite the increasing attention it attracts, the understanding of the characteristics and impact of remote peering is limited. In this work, we introduce and validate a heuristic methodology for discovering remote peers at IXPs. We (i) identify critical remote peering inference challenges, (ii) infer remote peers with high accuracy (>95%) and coverage (93%) per IXP, and (iii) characterize different aspects of the remote peering ecosystem by applying our methodology to 30 large IXPs. We observe that remote peering is a significantly common practice in all the studied IXPs; for the largest IXPs, remote peers account for 40% of their member base. We also show that today, IXP growth is mainly driven by remote peering, which contributes two times more than local peering.

On the Latency Impact of Remote Peering

“…Available Data Limits Accuracy of Remote Peering Inferences. The current state-of-the-art methodology for inferring remote peering proposed by Giotsas et al [27] infers remote peering (1) through a reseller and/or (2) geographically distant from the IXP. The method combines delay measurements with additional features, such as port capacity and AS presence at colocation facilities; if an AS is not present in one of the feasible IXP facilities, their method infers the AS is remotely connected.…”

“…We used available ground truth ( §2.2) for four IXPs (LINX, PTT-SP, PTT-RJ, and PTT-CE) and applied their method to all interfaces connected to these IXPs. We implemented the four steps from the Giotsas et al [27] method. The first step (ping measurement campaign) measures the latency to IXP member interfaces from a VP within the IXP.…”

“…Fig. 3: Classification of interfaces we obtained when we applied our implementation of the current state-of-the-art methodology for inferring remote peering [27]. The high percentage of no inference for the three Brazilian IXPs was a consequence of the method's high reliance on public information (PeeringDB) which was not widely available for members of Brazilian IXPs.…”

Internet interconnection infrastructure: lessons from the global South

“…4 Thus, a network needs to pay for transport to physically arrive at the data centre, e.g., hiring a fibre link 4. Recently, remote peering has created different dynamics for physical interconnection, especially in the global North, but this is not covered here (See Giotsas et al, 2021).…”

BGP-Multipath Routing in the Internet

“…Two peering ASes are neighbouring ASes if they are connected directly, via physical links or an IXP (for reduced latency and improved routing performance), and they are called neighbour ASes to each other. Otherwise, the two peering ASes are called remote ASes to each other (for monetary savings and increased connectivity [11]. An AS can deploy BGP-M to both neighbour ASes and remote ASes.…”