ChainSync: A Synchronization Protocol for Strict Sequential Execution in Linear Distributed Pipelines
draft-dohmeyer-chainsync-01
This document is an Internet-Draft (I-D).
Anyone may submit an I-D to the IETF.
This I-D is not endorsed by the IETF and has no formal standing in the
IETF standards process.
The information below is for an old version of the document.
| Document | Type |
This is an older version of an Internet-Draft whose latest revision state is "Active".
|
|
|---|---|---|---|
| Author | Douglas Russell Dohmeyer | ||
| Last updated | 2025-12-04 (Latest revision 2025-12-02) | ||
| RFC stream | (None) | ||
| Formats | |||
| Stream | Stream state | (No stream defined) | |
| Consensus boilerplate | Unknown | ||
| RFC Editor Note | (None) | ||
| IESG | IESG state | I-D Exists | |
| Telechat date | (None) | ||
| Responsible AD | (None) | ||
| Send notices to | (None) |
draft-dohmeyer-chainsync-01
Independent Submission Douglas Dohmeyer
Internet-Draft Independent Researcher
Intended status: Informational 5 December 2025
Expires: 8 June 2026
ChainSync: A Synchronization Protocol for Strict Sequential Execution in
Linear Distributed Pipelines
draft-dohmeyer-chainsync-01
Abstract
ChainSync is a lightweight application-layer protocol that runs over
reliable TCP connections to synchronize a fixed linear chain of
distributed processes such that they execute their local tasks in
strict sequential order and only after every process in the chain has
confirmed it is ready. The protocol has four phases: 1) a forward
"readiness" wave, 2) a backward "start" wave, 3) a forward
"execution" wave, and 4) a backward exit wave.
The design guarantees strict ordering even when nodes become ready at
very different times and requires only point-to-point TCP connections
along the chain, thus no central coordinator is needed.
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 8 June 2026.
Copyright Notice
Copyright (c) 2025 IETF Trust and the persons identified as the
document authors. All rights reserved.
Douglas Dohmeyer Expires 8 June 2026 [Page 1]
Internet-Draft ChainSync December 2025
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents (https://trustee.ietf.org/
license-info) in effect on the date of publication of this document.
Please review these documents carefully, as they describe your rights
and restrictions with respect to this document. Code Components
extracted from this document must include Revised BSD License text as
described in Section 4.e of the Trust Legal Provisions and are
provided without warranty as described in the Revised BSD License.
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1. Requirements Language . . . . . . . . . . . . . . . . . . 3
1.2. Topology and Configuration . . . . . . . . . . . . . . . 3
1.3. States . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4. Message Types . . . . . . . . . . . . . . . . . . . . . . 4
1.5. Protocol Operation . . . . . . . . . . . . . . . . . . . 4
1.5.1. Phase 1 -- Readiness Collection (Forward Wave) . . . 4
1.5.2. Phase 2 -- Start Trigger Propagation (Backward
Wave) . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5.3. Phase 3 -- Ordered Execution and Completion (Forward
Wave) . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5.4. Phase 4 -- Backward Propagating Exit (Backward
Wave) . . . . . . . . . . . . . . . . . . . . . . . . 5
1.6. Waiting in WATCH State . . . . . . . . . . . . . . . . . 5
1.7. Example Message Flow (A-B-C-D) . . . . . . . . . . . . . 5
2. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 6
3. Security Considerations . . . . . . . . . . . . . . . . . . . 6
4. Normative References . . . . . . . . . . . . . . . . . . . . 6
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . 6
Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 7
1. Introduction
Many distributed workflows (pipeline parallelism in machine-learning
training, staged data processing, multi-organization business
processes, ordered multi-phase computation, etc.) require that tasks
execute in a fixed order across different machines, yet must not
begin until every participant is ready.
Standard barriers do not enforce execution order. Token-passing or
leader-based schemes introduce complexity and single points of
failure.
ChainSync solves this with a simple, fully decentralized four-wave
algorithm on a line topology that guarantees:
1. No process starts until the entire chain is ready.
Douglas Dohmeyer Expires 8 June 2026 [Page 2]
Internet-Draft ChainSync December 2025
2. Execution order is strictly A -> B -> ... -> N.
3. Clean backward-propagating exit after N finishes.
1.1. 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.
1.2. Topology and Configuration
The processes form a static logical chain:
(Head) A <-> B <-> C <-> ... <-> N (Tail)
Each process knows:
* The IP address and port of its predecessor (Head has none)
* The IP address and port of its successor (Tail has none)
* Whether it is Head, Tail, or intermediate (inferable from the
presence/absence of a predecessor/successor)
Each adjacent pair maintains a single persistent bidirectional TCP
connection.
1.3. States
+==========+=======================================+
| State | Meaning |
+==========+=======================================+
| SYNC | Initial state; waiting for READY from |
| | the predecessor (Head starts here but |
| | moves to READY when locally ready) |
+----------+---------------------------------------+
| READY | Chain segment to the left is ready; |
| | has sent READY to the successor (if |
| | not Tail) |
+----------+---------------------------------------+
| WATCH | Has propagated START leftward; |
| | waiting for COMPLETE from the |
| | predecessor (if not Head) |
+----------+---------------------------------------+
| START | Currently executing its local task |
Douglas Dohmeyer Expires 8 June 2026 [Page 3]
Internet-Draft ChainSync December 2025
+----------+---------------------------------------+
| COMPLETE | Local task finished; has sent |
| | COMPLETE to its successor |
+----------+---------------------------------------+
Table 1
1.4. Message Types
Messages are simple ASCII text lines terminated by LF. Recommended
format:
<COMMAND>[:<ROUND-ID>]\n
Defined commands:
* READY[:<ROUND-ID>]
* START[:<ROUND-ID>]
* COMPLETE[:<ROUND-ID>]
<ROUND-ID> is optional but RECOMMENDED (e.g., UUID) to support
multiple concurrent rounds on the same connection. Implementations
running only one round at a time MAY omit it.
1.5. Protocol Operation
1.5.1. Phase 1 -- Readiness Collection (Forward Wave)
* Head (A), when locally ready, moves from SYNC to READY and sends
READY to its successor.
* Every other node starts in SYNC. When it receives READY from
predecessor *and* becomes locally ready, it moves from SYNC to
READY and sends READY to successor.
* When Tail (N) enters READY, Phase 2 begins automatically.
1.5.2. Phase 2 -- Start Trigger Propagation (Backward Wave)
* Tail, upon entering READY, sends START to its predecessor and
moves to WATCH.
* An intermediate node, upon receiving START from its successor:
1. Sends START to its predecessor
Douglas Dohmeyer Expires 8 June 2026 [Page 4]
Internet-Draft ChainSync December 2025
2. Moves to WATCH and waits for COMPLETE from its predecessor
* Head, upon receiving START, has no predecessor and therefore moves
directly to START and begins execution.
This phase completes in O(n) messages and guarantees every node knows
the entire chain is ready before any node starts.
1.5.3. Phase 3 -- Ordered Execution and Completion (Forward Wave)
* A node in WATCH that receives COMPLETE from its predecessor moves
to START and begins execution.
* When a node finishes its task, it moves from START to COMPLETE and
sends COMPLETE to its successor (triggers successor to start)
Execution order is therefore strictly A -> B -> C -> ... -> N.
1.5.4. Phase 4 -- Backward Propagating Exit (Backward Wave)
* Tail, upon entering COMPLETE has no successor and therefore
immediately sends COMPLETE to its predecessor and MAY terminate.
* An intermediate node in COMPLETE that receives COMPLETE from its
successor sends COMPLETE to its predecessor and MAY terminate.
* Head, upon receiving COMPLETE from its successor MAY terminate.
The completion of this phase guarantees the Head node knows all nodes
have completed execution.
1.6. Waiting in WATCH State
The RECOMMENDED approach is *push-based*: the node simply blocks on
read() from the predecessor's TCP socket. When the predecessor
finishes, it pushes COMPLETE. An alternative approach is to poll the
predecessor's TCP socket.
Both approaches are compliant.
1.7. Example Message Flow (A-B-C-D)
RD: READY
ST: START
CM: COMPLETE
Douglas Dohmeyer Expires 8 June 2026 [Page 5]
Internet-Draft ChainSync December 2025
A.....B.....C.....D
|-RD->|.....|.....| Phase 1
|.....|-RD->|.....|
|.....|.....|-RD->|
|.....|.....|<-ST-| Phase 2
|.....|<-ST-|.....|
|<-ST-|.....|.....| Phase 3
|.....|.....|.....| A starts immediately
|-CM->|.....|.....| A finishes and B starts
|.....|-CM->|.....| B finishes and C starts
|.....|.....|-CM->| C finishes and D starts
|.....|.....|.....| Phase 4
|.....|.....|<-CM-| D finishes
|.....|<-CM-|.....X D exits
|<-CM-|.....X...... C exits
|.....X............ B exits
X.................. A exits
2. IANA Considerations
This memo includes no request to IANA.
3. Security Considerations
Connections SHOULD use TLS 1.3. Production deployments SHOULD use
mutual TLS with certificate pinning or pre-shared keys to prevent
node impersonation.
4. Normative References
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119,
DOI 10.17487/RFC2119, March 1997,
<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,
May 2017, <https://www.rfc-editor.org/rfc/rfc8174>.
Acknowledgements
The author thanks Grok, an AI system developed by xAI, for assistance
in drafting portions of this document based on provided
specifications, for editing, for suggestions on the ROUND-ID
mechanism for concurrent rounds, and the backward propagation of
COMPLETE messages to ensure clean termination.
Douglas Dohmeyer Expires 8 June 2026 [Page 6]
Internet-Draft ChainSync December 2025
Author's Address
Douglas Russell Dohmeyer
Independent Researcher
United States of America
Email: douglas.dohmeyer@protonmail.com
Douglas Dohmeyer Expires 8 June 2026 [Page 7]