Can ISPs and P2P Users Cooperate for Improved Performance?
Vinay Aggarwal, Anja Feldmann Deutsche Telekom Laboratories/TU Berlin Berlin, Germany
Christian Scheideler TU München Munich, Germany
Peer-to-peer (P2P) systems, which are realized as overlays on top of the underlying Internet routing architecture, contribute a significant portion of today’s Internet traffic. While the P2P users are a good source of revenue for the Internet Service Providers (ISPs), the immense P2P traffic also poses a significant traffic engineering challenge to the ISPs. This is because P2P systems either implement their own routing in the overlay topology or may use a P2P routing underlay , both of which are largely independent of the Internet routing, and thus impedes the ISP’s traffic engineering capabil- ities. On the other hand, P2P users are primarily interested in finding their desired content quickly, with good performance. But as the P2P system has no access to the underlying network, it either has to measure the path per- formance itself or build its overlay topology agnostic of the underlay. This situation is disadvantageous for both the ISPs and the P2P users.
To overcome this, we propose and evaluate the feasibility of a solution where the ISP offers an “oracle” to the P2P users. When the P2P user supplies the oracle with a list of possible P2P neighbors, the oracle ranks them according to certain criteria, like their proximity to the user or higher bandwidth links. This can be used by the P2P user to choose appropriate neighbors, and therefore improve its performance. The ISP can use this mechanism to better manage the immense P2P traffic, e.g., to keep it inside its network, or to direct it along a desired path. The improved network utilization will also enable the ISP to provide better service to its customers.
Categories and Subject Descriptors
2.1 [Network Architecture and Design]: [Network topology];
2.4 [Distributed Systems]: [Distributed applications]
Design, Experimentation, Management, Performance
P2P, ISP, cooperation, routing, biased neighbor selection
P2P systems have recently gained a lot of attention from the In- ternet users and the research community. Popular applications that use P2P systems include file sharing systems such as Bit-torrent, eDonkey, Kazaa, Gnutella as well as VoIP systems such as Skype and GoogleTalk . P2P systems are so popular that they con- tribute more than 50% to the overall network traffic [3, 4, 5].
However, the wide-spread use of such P2P systems has put ISPs in a dilemma! On the one hand, P2P system applications have re- sulted in an increase in revenue for ISPs, as they are one of the ma- jor reasons cited by Internet users for upgrading their Internet ac- cess to broadband . On the other hand, ISPs find that P2P traffic poses a significant traffic engineering challenge [4, 7]. P2P traffic often starves other applications like Web traffic of bandwidth , and swamps the ISP network. This is because most P2P systems
rely on application layer routing based on an overlay topology on top of the Internet, which is largely independent of the Internet rout- ing and topology .
To construct an overlay topology, unstructured P2P networks usually employ an arbitrary neighbor selection procedure . This can result in a situation where a node in Frankfurt downloads a large content file from a node in Sydney, while the same informa- tion may be available at a node in Berlin. It has been shown that P2P traffic often crosses network boundaries multiple times [9, 10]. This is not necessarily optimal as most network bottlenecks in the Internet are assumed to be either in the access network or on the links between ISPs, but not in the backbones of the ISPs . Be- sides, studies have shown that the desired content is often available “in the proximity” of interested users [10, 12]. This is due to con- tent language and geographical regions of interest. Since a P2P user is primarily interested in finding his desired content quickly with good performance, we believe that increasing the locality of P2P traffic will benefit both ISPs and P2P users.
To better understand the origin of the problem of overlay-underlay routing clash, let us consider how routing works in the Internet and P2P systems. In the Internet, which is a collection of Autonomous Systems (ASes), packets are forwarded along a path on a per-prefix basis. This choice of path via the routing system is limited by the contractual agreements between ASes and the routing policy within the AS (usually shortest path routing based on a fixed per link cost) .
P2P systems, on the other hand, setup an overlay topology and implement their own routing  in the overlay topology which is no longer done on a per-prefix basis but rather on a query or key basis. In unstructured P2P networks queries are disseminated, e.g., via flooding  or random walks while structured P2P networks often use DHT-based routing systems to locate data . Answers can either be sent directly using the underlay routing  or through the overlay network by retracing the query path . By routing through the overlay of P2P nodes, P2P systems hope to use paths with better performance than those available via the Internet [14, 16]. But the benefits of redirecting traffic on an alternative path, e.g., one with larger available bandwidth or lower delay, are not necessarily obvious. While the performance of the P2P system may temporarily improve, the available bandwidth of the newly chosen path will deteriorate due to the traffic added to this path. The ISP then has to redirect some traffic so that other applications using this path receive enough bandwidth. In other words, P2P systems reinvent and reimplement a routing system whose dynamics should be able to interact with the dynamics of the Internet routing [7, 17]. While a routing underlay as proposed by Nakao et al.  can reduce the work duplications it cannot by itself overcome the interaction problems. Consider a situation where a P2P system imposes a lot of
ACM SIGCOMM Computer Communication Review
Volume 37, Number 3, July 2007