---
title: "Comment by Arnold Holm on \"A Scalable Integration Framework for Multi-Counterparty Derivatives and Financial Operations\" | Hashnode"
description: "I would make the ledger-posting crash window an explicit acceptance case: the ledger commits a balanced journal, but the consumer loses the acknowledgement and"
url: "https://hashnode.com/posts/a-scalable-integration-framework-for-multi-counterparty-derivatives-and-financial-operations/6ad31d92cca0bdd79faf9224/comment/6ad37a447cddfc4b2662ce51"
---

[Discussion](https://hashnode.com/posts/a-scalable-integration-framework-for-multi-counterparty-derivatives-and-financial-operations/6ad31d92cca0bdd79faf9224)

Comment [A Scalable Integration Framework for Multi-Counterparty Derivatives and Financial Operations](https://hashnode.com/posts/a-scalable-integration-framework-for-multi-counterparty-derivatives-and-financial-operations/6ad31d92cca0bdd79faf9224)

[Arnold Holm](https://hashnode.com/@stratcorealpha)

[Oct 17](https://hashnode.com/posts/a-scalable-integration-framework-for-multi-counterparty-derivatives-and-financial-operations/6ad31d92cca0bdd79faf9224/comment/6ad37a447cddfc4b2662ce51)

I would make the ledger-posting crash window an explicit acceptance case: the ledger commits a balanced journal, but the consumer loses the acknowledgement and replays the event. The replay should return the same posting reference and create no second journal. A transaction reference alone needs an atomic uniqueness check at the posting boundary, not only deduplication in the message consumer.
