Report from the IETF Workshop on Peer-to-Peer (P2P) Infrastructure, May 28, 2008

Jon Peterson, A. Feder Cooper2009 · DOI

TL;DR

This report summarizes an IETF workshop focused on the network congestion and delay issues caused by increasing Peer-to-Peer (P2P) traffic. It identifies key technical problems faced by ISPs and end-users, and proposes potential areas for IETF standardization to address these challenges, aiming for mutually beneficial solutions for network operators, application providers, and users.

What Problem Does This Paper Address?

The paper addresses the growing problem of network congestion and delay, particularly impacting real-time applications like Voice over IP (VoIP) and online gaming. This congestion is largely attributed to the high volume of traffic generated by Peer-to-Peer (P2P) applications. Traditional network management techniques are strained by applications designed to maximize end-user transfer rates, leading to near-absolute capacity utilization during peak times. The core challenge is to understand these technical issues from both ISP and end-user perspectives and to identify how the Internet Engineering Task Force (IETF) can contribute to engineering solutions that improve the overall Internet experience without stifling innovation.

Methods

This document is a report on a workshop, not a research paper with novel experimental methods. The "methods" employed were:

  • Workshop Organization: The IETF Real-time Applications and Infrastructure area directors organized a workshop to bring together subject matter experts.
  • Solicitation of Contributions: Position papers were solicited from participants prior to the workshop.
  • Expert Presentations: Invited speakers presented their views on the P2P infrastructure problem and potential solutions.
  • Discussion and Outcome Synthesis: The workshop facilitated discussions, and this report synthesizes the participants' contributions, identifying key problems and potential work items.
  • Categorization of Solutions: Proposed solutions were broadly categorized into improving peer selection, new approaches to congestion control, and quality-of-service mechanisms.

Datasets

Not discussed in the paper. The document reports on discussions and outcomes from a workshop, not on empirical studies or evaluations using specific datasets.

Key Insights and Ideas

  • P2P Traffic is a Systemic Issue: The problem isn't just about P2P itself, but about high-volume, always-on applications in general, and how they strain network resources beyond traditional traffic management.
  • ISP and Application Provider Perspectives Differ: ISPs face operational challenges and capacity constraints, while application providers focus on maximizing user experience through efficient data transfer, sometimes leading to increased network load.
  • Layer 2 Network Impact: The behavior of P2P traffic and congestion can be significantly influenced by lower-layer network technologies (e.g., DOCSIS in cable networks), which need to be considered in any solution.
  • Localization as a Key Solution Area: Improving peer selection by making applications aware of network topology, costs, and constraints (e.g., through ISP-provided information) holds significant promise for reducing network load and improving performance.
  • Need for Cooperative Frameworks: Solutions like P4P (Provider Portal for P2P Applications) aim to create a framework for cooperation between network providers and applications, allowing for more informed peer selection.
  • Congestion Control Evolution: Existing congestion control mechanisms (like TCP's) may need augmentation or new approaches (e.g., weighted congestion control, deferential end-to-end control) to better handle the characteristics of P2P traffic and interactive applications.
  • Quality of Service (QoS) Potential: While short-term fixes like Deep Packet Inspection (DPI) have limitations, standardized QoS mechanisms like DSCP could offer long-term benefits for traffic management and usage-based charging.
  • IETF's Role: The IETF's focus should be on standardizing core, manageable components that are not overly specific to any single ISP or P2P network, aiming for broad applicability.

Key Results

The workshop identified several key areas for potential IETF work:

  • Transport-Related Work Items:
    • Creation of a standardized, delay-based, end-to-end congestion-control mechanism for P2P clients to reduce their impact on interactive applications.
  • An informational document describing current practices of P2P applications opening multiple TCP connections and providing best-practice recommendations.
  • Improved Peer Selection:
    • Leveraging AS Numbers for more efficient peer selection.
  • Developing frameworks like P4P for information sharing between networks and applications.
  • Exploring multi-layer, tracker-based architectures for localization.
  • Investigating ISP-aided neighbor selection services.
  • Considering the role of caches in P2P networks.
  • New Approaches to Congestion Control:
    • Exploring end-to-end congestion control mechanisms that prioritize interactive traffic.
  • Investigating weighted congestion control for better bandwidth allocation.
  • Quality of Service (QoS):
    • Standardizing QoS mechanisms like DSCP for traffic classification and potential usage-based charging.

Limitations and Future Work

  • Data Scarcity: The paper highlights a need for more data on Internet congestion, its location, duration, frequency, effects, and root causes. Strategies for reliably collecting and disseminating this data are still elusive.
  • Incentive Alignment: The incentives for information sharing between ISPs and P2P clients are complex and need careful consideration to ensure mutual benefit.
  • Legal Liability: Concerns about legal liability, particularly regarding hosting caches or providing information that could facilitate the distribution of copyrighted content, may hinder adoption.
  • Scalability Challenges: Implementing solutions like ISP-hosted oracles faces scalability challenges due to the potential volume of requests from a large number of peers.
  • Future IETF Work: The report emphasizes that the IETF's task is to pinpoint incrementally beneficial work items, focusing on core, manageable components rather than full P2P mechanisms initially. Further research is needed to evaluate the effects of localization on various network and overlay architectures, and to understand swarm distributions. The report suggests that issues related to cost and congestion might be suitable for IRTF (Internet Research Task Force) work. Security considerations for any proposed mechanisms are also highlighted as crucial.

Summary updated

Abstract

This document reports the outcome of a workshop organized by the Real- time Applications and Infrastructure Area Directors of the IETF to discuss network delay and congestion issues resulting from increased P2P traffic volumes. The workshop was held on May 28, 2008 at MIT in Cambridge, MA, USA. The goals of the workshop were twofold: to understand the technical problems ISPs and end users are experiencing as a result of high volumes of P2P traffic, and to begin to understand how the IETF may be helpful in addressing these problems. Gaining an understanding of where in the IETF this work might be pursued and how to extract out feasible work items were highlighted as important tasks in pursuit of the latter goal. The workshop was very well attended and produced several work items that have since been taken up by members of the IETF community.
BibTeX
@misc{peterson2009report,
  author = {Jon Peterson and A. Feder Cooper},
  title = {{Report from the IETF Workshop on Peer-to-Peer (P2P) Infrastructure, May 28, 2008}},
  year = {2009},
  doi = {10.17487/rfc5594},
}

Loading related content...

Figure