Internet Engineering Task Force P. Crowley Internet Draft Washington University Intended status: Informational July 7, 2008 Expires: January 1, 2009 On the Relative Importance of P2P Peer Selection draft-crowley-alto-importance-01.txt Status of this Memo By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she becomes aware will be disclosed, in accordance with Section 6 of BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html This Internet-Draft will expire on January 1, 2009. Copyright Notice Copyright (C) The IETF Trust (2008). Abstract This Internet-draft discusses the relative importance of path selection in peer-to-peer (P2P) applications in light of the recent discussions highlighting the conflict between the use of P2P applications and the costs borne by network infrastructure operators. Crowley Expires January 1, 2009 [Page 1] Internet-Draft Relative Importance of Path Selection July 2008 Conventions used in this document The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC-2119 [1]. Table of Contents 1. Introduction...................................................2 2. Context & Importance...........................................2 2.1. Technical Strengths.......................................3 3. Problems with P2P Applications.................................4 4. Proposed Solutions.............................................5 5. Problems Solved via Standardization............................6 6. What is Needed.................................................6 7. Security Considerations........................................7 8. IANA Considerations............................................7 9. Conclusions....................................................7 10. Acknowledgments...............................................7 11. Informative References........................................7 Author's Addresses................................................8 Intellectual Property Statement...................................8 Disclaimer of Validity............................................8 1. Introduction The purpose of this document is to juxtapose peer-to-peer (P2P) application context, importance, and problems with solutions offered by standardization. In particular, the goal is to explore whether the current challenges created by P2P applications can be effectively solved, or at least ameliorated, through standardization efforts. 2. Context & Importance P2P file-sharing is a significant consumer Internet traffic source. A recent report from Cisco Systems [1,2] indicates that in 2007 P2P applications accounted for 51% of global consumer Internet traffic. While there are non-file-sharing P2P applications, file sharing is the dominant source of bandwidth consumption. P2P-based streaming video is significant in some parts of the world, but that is not counted among the P2P numbers in the Cisco report. We can conclude, therefore, that P2P file-sharing is an important application for users and infrastructure owners. However, it is also reasonable to assume that the majority of the bandwidth consumed by P2P applications is devoted to the distribution Crowley Expires January 1, 2009 [Page 2] Internet-Draft Relative Importance of Path Selection July 2008 of video files. While some may assume that the popularity of P2P applications is content-independent, it is reasonable to suppose that the popularity of P2P applications, the least when measured by bandwidth consumed, is due entirely to the popularity of sharing video files. Anecdotally, it seems that this file-sharing is mostly illegal; whether this is blatant theft or a form of civil disobedience is a topic left for another time. More to the point, the Cisco study also reports that non-P2P video traffic has experienced, and will continue to experience, an even greater rate of growth than P2P applications. Indeed, despite accounting for 51% of global Internet traffic, P2P traffic shrank year-on-year as a total percentage while non-P2P video grew. Before discussing the technical strengths and problems with P2P applications, we note that this is an unwieldy conversation because many topics are customarily grouped together, while specific problems and solutions typically apply only to a single topic. Topics that are usually aggregated included: o BitTorrent and the characteristics that distinguish it from other P2P clients and protocols. o File Sharing as opposed to other applications on peer-to-peer platforms, such as video-on-demand, voice and instant messaging. o P2P versus client-server service architectures. In this Internet-draft, we will focus the discussion on the use of P2P applications for file-sharing, because this usage dominates both the importance and the popularity of P2P systems in the world today. 2.1. Technical Strengths P2P systems, and in particular BitTorrent, are in most ways remarkably efficient technologies. BitTorrent effectively eliminates costs associated with distributing large digital files. Individuals with consumer DSL connections can be global-scale publishers of multi-gigabyte files. It is for this reason that people around the world can rip and share extremely large video files at very little cost. P2P systems deploy and evolve at the rate of change of end-user habits. There is very low friction resisting development and deployment. There are no encumbrances from service or infrastructure providers to slow things down. Crowley Expires January 1, 2009 [Page 3] Internet-Draft Relative Importance of Path Selection July 2008 Perhaps most remarkably, file distribution with BitTorrent scales automatically with available resources, because it does not require manual intervention to utilize capacity increases. 3. Problems with P2P Applications The problems caused by P2P applications can be categorized according to the party experiencing the problem. Home users, those who install and use P2P clients on their own computers, experience at least two kinds of problems. First, P2P file-sharing applications interfere with latency sensitive applications, such as voice over IP (VOIP), because they tend to fill all available buffers between a computer and the end of its access link. As recently pointed out by S. Shalunov at the IETF P2P infrastructure workshop in May 2008, mass-market consumer access points and switches feature small amounts of buffer space that BitTorrent tends to fill, increasing ping times and jitter for latency sensitive applications. Second, P2P applications can easily exhaust bandwidth caps associated with broadband connections. For ISPs and network infrastructure operators, P2P applications present at least three problems: large bandwidth usage, usage that scales with available capacity, and steady, long-lived connections rather than bursty ones. Since files shared on P2P networks tend to be large, the aggregate bandwidth used is also large. As mentioned above, a P2P application such as BitTorrent automatically scales to take advantage of the available network resources. This creates a tremendous problem for increasing capacity: if rising P2P traffic has inspired an accelerated capacity increase and that P2P traffic can automatically scale to take advantage of any added capacity, then it is difficult to make rational, cost-effective capacity investment plans. Long-lived, non-bursty traffic is troublesome because it runs afoul of the assumptions that govern the oversubscription rates used to price broadband connections. It can be seen that the problem is in fact due to the conjunction of these three. For example, FTP is also steady, large, and long-lasting but very few use it so it causes no large problem. It is worth asking whether these are technical problems or business model problems. Certainly, a technical decision could be taken to Crowley Expires January 1, 2009 [Page 4] Internet-Draft Relative Importance of Path Selection July 2008 remove buffering from access points and switches, although this would not apply to the installed base of tens of millions of devices. However, the other problems are at least equally dependent on business model or social factors, rather than on solely technical questions. 4. Proposed Solutions A high-level notion of ISP cost is the problem most discussed in popular media and research literature. Many proposals for P2P peer- localization, content caching, and hybrid server-peer architectures have been proposed. Additionally, commercial offerings include devices that can measure P2P traffic and apply traffic shaping policies. ALTO is particularly concerned with the mechanisms used to select peers, which in turn can achieve localization or effective caching. Examples of these mechanisms are nicely surveyed in the ALTO abstract presented at the IETF P2P infrastructure workshop in May 2008, where two major approaches are summarized: end-to-end measurement and application-to-network layer cooperation. In general, however, these peer selection ideas are only useful in contexts where peers are abundant and many options of equivalent quality exist. Whenever peers must be chosen based on the rare content they can provide, peer selection mechanisms will not apply. Of course, it is natural to assume that a disproportionately large amount of P2P file sharing bandwidth can be accounted for by a relatively small number of distinct files. That is, a subset of popular videos will likely account for the majority of traffic at any given time. However, as other legal online video options proliferate, it is possible that P2P file-sharing will cease to be the primary video distribution platform. In recent years, the number of ad-supported and subscription-based online video services has flourished. Continued developments along this path-along with the possibility that the public-at-large may eventually lose its appetite for illegal file sharing-may significantly alter P2P dynamics. And more to the point, as video continues to dominate global Internet traffic, the Internet will experience dramatic increases in the number of high-bandwidth, non-bursty, long-lived connections. P2P applications are singled out as the problematic ones today, but in the near future there will be many more video-oriented applications that will cause similar problems. Crowley Expires January 1, 2009 [Page 5] Internet-Draft Relative Importance of Path Selection July 2008 5. Problems Solved via Standardization What Problems do Standards Solve? In general, standards have two effects. They avoid the problems associated with a single solution provider, such as monopoly pricing power and technical frailty, by encouraging commoditization of the object of standardization. They also avoid the problems associated with the reinvention and fragmentation of common solutions, such as: safety (i.e., standards can reduce the likelihood of errors and mistakes), interoperability, and quality. 6. What is Needed There is no shortage of feasible technical solutions to the problems raised by P2P applications, but the effectiveness of a given solution will be determined by business models and end-user willingness to adopt them. Examples of connectivity business models include the following. o Fixed prices with bandwidth caps. o Pay-per-bit. o Traffic categories, with per-category pricing models. The current user vocabulary defines as categories phone, video, and data. This is obvious and clear to consumer customers of telephone and cable companies. P2P traffic characteristics cause a problem for existing data pricing models. Some have suggested that data pricing models require further elaboration, with finer subdivisions. For example, one can imagine web/e-mail/data, P2P, gaming, voice, and video all be as coming subcategories within a data plan, each with their own pricing model. Might users understand and accept such a pricing model? Would such a model make sense for broadband providers? This line of thinking leads one to ask: Is the Internet a utility, an infrastructure, or a service platform? If it is a utility, then we can consider the pricing model for electricity, in which users are charged for each kilowatt-hour used, without regard for how the units are used (but perhaps with regard to when they are used). Crowley Expires January 1, 2009 [Page 6] Internet-Draft Relative Importance of Path Selection July 2008 If it is an infrastructure, then we can consider the pricing model for interstate highways. Such a model is very different from today's, since this type of infrastructure does not produce profit directly, but enables profitable uses. If the Internet is a service platform, such as mobile phones and mobile phone networks, then one can imagine pricing models that separate out distinct uses, such as with mobile voice, data, and SMS pricing models. In such a model, not all bits carry the same price. 7. Security Considerations Security risks tend to increase with complexity. Any proposal that increases the complexity of infrastructure or its interfaces, such as those that aim to provide access to direct or indirect topology information, invites the risk of complicating security. More specifically, if a service provider publishes information that is meant to both improve end-user performance and reduce service provider cost, there will always be the opportunity for subversion, or perhaps the more insidious suspicion that the service provider is paying greater attention to its costs than to the performance provided. An effective antidote for this situation is independent verification. 8. IANA Considerations 9. Conclusions In conclusion, because standardization seems ill-positioned to address the most pressing problems with P2P systems, it is likely premature for the IETF to pursue any strong activities in standardization. There may be other reasons to pursue standardization, but there does not appear to be a strong, objective case for doing so for the sake of improving P2P file-sharing. 10. Acknowledgments This document was prepared using 2-Word-v2.0.template.dot. 11. Informative References [1] Cisco Systems, "Cisco Visual Networking Index-Forecast and Methodology, 2007-2012", White Paper, http://www.cisco.com, June 2008. Crowley Expires January 1, 2009 [Page 7] Internet-Draft Relative Importance of Path Selection July 2008 [2] Cisco Systems, "Approaching the Zettabyte Era", White Paper, http://www.cisco.com, June 2008. Author's Addresses Patrick Crowley Washington University Dept. of CSE/Campus Box 1045 One Brookings Drive St. Louis, MO 63130 USA Email: pcrowley@wustl.edu URL: http://www.arl.wustl.edu/~pcrowley Intellectual Property Statement The IETF takes no position regarding the validity or scope of any Intellectual Property Rights or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; nor does it represent that it has made any independent effort to identify any such rights. Information on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79. Copies of IPR disclosures made to the IETF Secretariat and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementers or users of this specification can be obtained from the IETF on-line IPR repository at http://www.ietf.org/ipr. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at ietf-ipr@ietf.org. Disclaimer of Validity This document and the information contained herein are provided on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF Crowley Expires January 1, 2009 [Page 8] Internet-Draft Relative Importance of Path Selection July 2008 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Copyright Statement Copyright (C) The IETF Trust (2008). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. Acknowledgment Funding for the RFC Editor function is currently provided by the Internet Society. Crowley Expires January 1, 2009 [Page 9]