Dynamic Partitioning for Partially Replicated Databases

部分复制数据库的动态分区

• 批准号: 543858-2019
• 负责人: Daudjee, Khuzaima
• 金额: $ 5.77万
• 依托单位: University of Waterloo
• 依托单位国家: 加拿大
• 项目类别: Collaborative Research and Development Grants
• 财政年份: 2021
• 资助国家: 加拿大
• 起止时间: 2021-01-01 至 2022-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=720805
• 关键词: Dynamic; Partitioning; Partially; Replicated; Databases

To scale-up data services, database systems popularly leverage replication or partitioning to increase server resources by distributing data partitions and their replicas over multiple servers. Data partitioning and replication become challenging tasks in the face of changing workload access patterns, which result in load imbalance and performance degradation. We propose to design and build DRP, a system that will partition and replicate data dynamically to account for changes in the workload. DRP will deliver superior performance by mitigating the effects of load skew and contention that arise from a shifting and changing workload.

为了扩展数据服务，数据库系统普遍利用复制或分区来通过将数据分区及其副本分布在多个服务器上来增加服务器资源。面对不断变化的工作负载访问模式，数据分区和复制成为具有挑战性的任务，从而导致负载不平衡和性能下降。 我们建议设计和构建DRP，这是一个动态分区和复制数据以应对工作负载变化的系统。 DRP将通过减轻负载偏移和争用的影响来提供上级性能，而负载偏移和争用是由于工作负载的变化而引起的。