Statistical Quality of Service Guarantee for Temporal Consistency of Real-Time Data Objects

The More-Less (ML) scheme is an efficient technique for maintaining temporal consistency of real-time data objects [13]. ML provides deterministic guarantee in temporal consistency. However, the number of transactions that can be supported using ML is severely limited due to the use of worst-case computation time of transactions. In this paper, we study the problem of temporal consistency maintenance for a real-time database system where a certain degree of temporal inconsistency is tolerable. We propose a series of approaches that enable the tradeoff of quality of service (QoS) of temporal consistency against the number of supported transactions. We start with a base-line algorithm, Statistical More-Less-Base-line (SML-BA), that provides the requested QoS of temporal consistency. We then propose SML with Optimization, namely SML-OPT, that further improves the QoS by better utilizing the excessive CPU capacity. Finally, SML-OPT is enhanced with a Slack Reclaiming scheme (SML-SR).