{"title":"Bilateral anti-entropy for eventual consistency","authors":"Rebecca Bilbro, Benjamin Bengfort, P. Keleher","doi":"10.1145/3517209.3524083","DOIUrl":null,"url":null,"abstract":"Eventually consistent systems are often more cost-effective to implement and maintain than their strongly consistent cousins. Gossip-based anti-entropy methods can be used to improve the consistency of such systems. However, observational data suggests that such improvements come at the cost of perennial data egress, significantly mitigating savings in a geo-replicated context. This paper presents observations collected from an eventually consistent system deployed in production clusters across the US, Germany, and Singapore. A bilateral anti-entropy process facilitates data synchronization for an application accessed globally and around-the-clock. Replication metrics such as latency, conflict, and the duration and productivity of anti-entropy sessions seem to indicate that it is possible to quantify tradeoffs between consistency and visibility latency such that staleness can be not only bounded, but fine-tuned.","PeriodicalId":445734,"journal":{"name":"Proceedings of the 9th Workshop on Principles and Practice of Consistency for Distributed Data","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 9th Workshop on Principles and Practice of Consistency for Distributed Data","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3517209.3524083","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Eventually consistent systems are often more cost-effective to implement and maintain than their strongly consistent cousins. Gossip-based anti-entropy methods can be used to improve the consistency of such systems. However, observational data suggests that such improvements come at the cost of perennial data egress, significantly mitigating savings in a geo-replicated context. This paper presents observations collected from an eventually consistent system deployed in production clusters across the US, Germany, and Singapore. A bilateral anti-entropy process facilitates data synchronization for an application accessed globally and around-the-clock. Replication metrics such as latency, conflict, and the duration and productivity of anti-entropy sessions seem to indicate that it is possible to quantify tradeoffs between consistency and visibility latency such that staleness can be not only bounded, but fine-tuned.