6. AQUAS: A Quality-Aware Scheduler

In order to study the role of quality-aware scheduling in practice, we introduce a quality-aware scheduler named AQUAS (Xu et al., AQUAS: a quality-aware scheduler for NoSQL data stores 1210–1213, 2014 [1]) which has been integrated into Cassandra. This prototype employs the strategies proposed in previous chapters to manage the quality-aware scheduling in key-value stores. In particular, AQUAS extends the Cassandra API to allow application developers to specify quality-aware parameters, and implements a task scheduling component as well as a cost estimation component. In addition, we propose a modified YCSB (Coopre et al., Benchmarking cloud serving systems with YCSB 143–154, 2010 [2]) to evaluate AQUAS prototype, and explore the timeline query in miroblogging application to demonstrate the role of quality-aware scheduling based on AQUAS prototype. In the following, Sect. 6.1 describes the design of a quality-aware scheduler named AQUAS; Sect. 6.2 illustrates a benchmark named QCSB, which is tailored for AQUAS, as well as the performance study; Sect. 6.3 takes the timeline query on Microblogging application as a example, to depict the application supported by AQUAS; Sect. 6.4 summaries this chapter.