Internet-Draft Precise Priority-based Flow Control with June 2026
Xiong & Zhu Expires 17 December 2026 [Page]
Workgroup:
rtgwg
Internet-Draft:
draft-xz-rtgwg-ppfc-gateway-00
Published:
Intended Status:
Standards Track
Expires:
Authors:
Q. Xiong
ZTE Corporation
X. Zhu
ZTE Corporation

Precise Priority-based Flow Control with Gateway

Abstract

This document proposes Precise Priority-based Flow Control (PPFC) mechanism implemented at the network gateway, it designed to efficiently manage congestion in scenarios where multiple flows converge toward a common destination. By enabling fast, per-flow congestion control at the network gateway, PPFC allows incoming traffic destined for a congested bottleneck to be paused or rate-limited before entering the network. This approach decouples the congestion feedback loop from the end-to-end data path, significantly reducing reaction time and improving the effectiveness of congestion management.

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.

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."

This Internet-Draft will expire on 17 December 2026.

Table of Contents

1. Introduction

In modern network deployments such as data centers and wide-area interconnections, multiple source clients frequently communicate through a shared ingress gateway. When substantial traffic is directed to a common server, these flows often converge onto a single bottleneck link within the network. Traditional end-to-end congestion control mechanisms��such as BBR and ECN��rely on feedback that must traverse the full return path to senders, introducing at least one round-trip time (RTT) of delay before rate adjustments take effect. This delay is particularly problematic in high-throughput, low-latency environments where transient congestion can severely impact performance.

The PPFC (Precise Priority-based Flow Control) as per [I-D.xz-rtgwg-ppfc-notification] can address this limitation by enabling network-assisted congestion control near the sources clients. The key idea is to instantaneously pause or throttle sources at the ingress point (e.g., a gateway), allowing the network to quickly recover from congestion.

This document proposes an enhancement to existing congestion control by introducing a PPFC mechanism at the gateway. By deploying PPFC at a gateway that aggregates traffic from multiple sources destined for a common destination, the gateway can prevent new flows from exacerbating existing congestion. It will significantly mitigate the congested link when the traffic transfers to the same destination.

2. Conventions Used in This Document

2.1. Abbreviations

PPFC: Precise Priority-based Flow Control

RTT: Round-Trip Time

TCP: Transfer Control Protocol

RDMA: Remote Direct Memory Access Round-Trip Time

QUIC: Quick UDP Internet Connections

ECN: Explicit Congestion Notification

PFC: Priority-based Flow Control

RoCEv2: RDMA over Converged Ethernet version 2

2.2. Requirements Language

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.

3. Precise Priority-based Flow Control with Gateway

Upon receiving the congestion in PPFC mechanism, the gateway can record the states and apply immediate congestion control to any subsequent traffic for that server. Furthermore, the gateway can perform source flow control, when new traffic arrives, it may reject or pause new flows to prevent worsening existing congestion.

The following figure illustrates this interaction:

               3-PPFC Notification (e.g.Server, Pause, 10us)
         ***************************************************
         *                     *                           *
         *                     *                           *
         *                     *                           *
         V                     V                           *
    +---+----+ 1-Traffic A +---+---+     +-------+     +---+---+     +-------+
    |Client A|<----------->|Gateway|<--->|Node X |<--->|Node Y |<--->|Server |
    +--------+      +----->+-------+     +-------+     +-------+     +-------+
         4-Traffic B|  ****+                           2-Congestion
    +--------+      |  *                                 Occurs
    |Client B+<-----+  * 5-PPFC Fast Feedback (e.g.Server, Pause, 10us)
    +--------+<*********

                         Figure 1: PPFC Mechanism at Gateway

The example at gateway is shown in Figure 1 and the steps are as follows.

Step 1: Client A sends traffic to the Server via the Gateway, Node X, and Node Y.

Step 2: Congestion occurs at the interface of Node Y towards the Server.

Step 3: Node Y generates an PPFC Notification containing the destination (Server) and a pause duration (10us), sending it to the Node X and the Node X transits it to the the Gateway.

Step 4: Client B attempts to send traffic to the same Server.

Step 5: The Gateway, having recorded the congested state for the Server, applies source flow control to Client B's new traffic. This may involve holding packets or sending an PPFC Feedback message with pause duration (e.g. 10us). which may apply the format as per [I-D.xz-rtgwg-ppfc-rocev2] to Client B, preventing its traffic from entering the network and aggravating the congestion at Node Y.

4. Security Considerations

To be discussed in future versions of this document.

5. IANA Considerations

This document does not currently require any IANA actions.

6. Contributors

The following people have substantially contributed to this document:


Qinghua Shao
ZTE Corporation
Email: shao.qinghua@zte.com.cn

7. Normative References

[I-D.xz-rtgwg-ppfc-notification]
Xiong, Q. and X. Zhu, "Precise Priority-based Flow Control Notification", Work in Progress, Internet-Draft, draft-xz-rtgwg-ppfc-notification-00, , <https://datatracker.ietf.org/doc/html/draft-xz-rtgwg-ppfc-notification-00>.
[I-D.xz-rtgwg-ppfc-rocev2]
"*** BROKEN REFERENCE ***".
[RFC2119]
Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, , <https://www.rfc-editor.org/rfc/rfc2119>.
[RFC8174]
Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, , <https://www.rfc-editor.org/rfc/rfc8174>.

Authors' Addresses

Quan Xiong
ZTE Corporation
Xiangyang Zhu
ZTE Corporation