Towards enhancing the quality of real-time data services

It is essential to process real-time data service requests such as traffic status queries and trade transactions in a timely manner using fresh data, which represent the current real world status. However, it is challenging to support real-time data services, if workloads dynamically change based on the market or traffic status.;To enhance the quality of real-time data services, we develop (1) a fine-grained closed-loop admission control scheme based on our novel backlog model; (2) a feed-forward approach that proactively adapts the incoming load to support the desired real-time data service delay and integrate this proactive approach with a feedback controller to compensate potential prediction errors in a reactive manner for timely data services; (3) a new deadline assignment scheme to derive feasible deadlines for real-time data service requests considering their individual data needs; (4) a systematic closed-loop approach to supporting the desired tardiness---the actual service delay to deadline ratio---of real-time data services; (5) a new approach to assigning feasible differentiated deadlines to real-time data service requests considering both individual data needs and service classes. Our approach assigns the shortest (longest) relative deadline to the smallest (largest) request belonging to the highest (lowest) service class, while trying to meet as many deadlines of the admitted requests as possible by closely supporting the desired tardiness via feedback control.;Performance evaluation results acquired in our real-time data service testbed show that our closed-loop admission control schemes considerably outperform several baselines representing the current state of art in terms of the data service delay and throughput. Our feed-forward and feedback approaches effectively reduces the average delay and transient delay fluctuations, while improving throughput compared to the baselines including the feed-forward-only and feedback-only approaches. Also, the proposed tardiness control approaches can closely support the desired average/transient tardiness. Consequently, the deadline miss ratio is significantly reduced compared to a state-of-art database system with a real-time scheduling extension. The experimental results of our differentiated deadline assignment and tardiness control approach show that our real-time database system closely supports the desired tardiness, while differentiating the delay across the data service classes.