紀要論文

双方向変換網における操作変換を用いたデータ更新競合解決手法

Operational Transformation-based Conflict Resolution for Data Updates in Bidirectional Transformation Network

jpn

24321192

https://doi.org/10.15002/00030619

法政大学大学院情報科学研究科

法政大学大学院紀要. 情報科学研究科編

19

1

9

2024-03-24

Version of Record

Bidirectional transformation enables data synchronization between multiple sources, but handling multiple conflicting updates poses a challenge. Habu et al. solve this problem using an operational transformation, but it is based on the convergence property C_1, which assumes that only two updates conflict. When a network of bidirectional transformations is used to synchronize more than two sources, more than two updates may conflict. In such cases, the convergence property C_2 is also required. However, the normal API which is based on inserts and deletes, cannot be used since it necessitates an additional parameter to the update API. To ensure confluence with more than two updates for the normal API (and thus without C_2), appropriate control is needed when operational transformations are composed. This study proposes a solution for handling multiple conflicting updates in a network of bidirectional transformations using controlled operational transformations. Our extension to the duplicate primitive “Dup” enables a 4-way merge in its backward transformation. We use Orderer, a mediator of the distributed ledger platform Hyperledger Fabric, to achieve the aforementioned control. Our implementation assumes a collaborative editing for XML via three views and we confirmed through experiments the well-behavedness of the bidirectional transformation including duplicates and that update intentions were preserved in the final result.

Article

AA12746425