| The demand for real-time data services is increasing in many applications including e-commerce, agile manufacturing, and traffic control. In these applications, it is essential to execute transactions within their trade or process control deadlines using fresh (temporally consistent) data that reflect the current market or manufacturing state. However, meeting these fundamental requirements is challenging because workloads and data access patterns may vary dynamically. Further, timeliness and data freshness requirements may conflict.; To address this problem, I apply QoS management techniques to manage the performance of real-time databases. My approach, called QMF (a QoS management architecture for Miss ratio and Freshness guarantees), provides several QoS metrics to let a database administrator specify the required deadline miss ratio and data freshness for a specific application. Unlike non-real-time/QoS-unaware approaches. QMF can support the required QoS for admitted transactions even in the presence of unpredictable workloads and data access patterns. In this way, QMF can significantly decrease the potential loss of profit or product quality due to timing and freshness constraint violations prevalent in dynamic environments. I also present novel approaches for service differentiation and secure transaction processing in real-time databases, called QMF-Diff (QMF with Differentiated services) and STAR (Security and Timeliness Assurance in Real-time databases), respectively. Using QMF-Diff, real-time transactions can be classified into several service classes according to their importance, and receive differentiated miss ratio guarantees. Therefore, limited system resources can be effectively utilized especially when overloaded. STAR is novel in that it does not only support security assurance, but also gives real-time performance guarantees. Consequently, STAR can protect, for example, sensitive business or military data, while processing transactions in a timely manner. |
