tailieunhanh - Misconceptions About Real-Time Databases

We show experiments for on-line convergence to a global partitioning solution for sharing the database buffer pool, storage cache, and disk bandwidth in dif- ferent application configurations. We compare our ap- proach to two baseline approaches, which optimize ei- ther the memory partitioning, or the disk partitioning, as well as combinations of these approaches without global coordination. We show that for most application con- figurations, our computed model effectively prunes most of the search space, even without any additional tuning through experimental sampling. Our dynamic resource algorithmperforms similar to an experimental exhaustive search algorithm, but provides a solution within minutes, versus days of running time. At the same time, our global resource partitioning solution. | Cybersquare John A. Stankovic and Sang Hyuk Son University of Virginia Charlottesville Jorgen Hansson University of Skovde Sweden Misconceptions About Real-Time Databases D atabases have become an integral part of many computer systems ranging from complex systems that control air traffic plant operations and stock mar- Some database users think real-time1 databases just need to be fast and that conventional databases are adequate for real-time applications. Real-time database designers don t agree. In a real-time database system timing constraints are associated with transactions and data are valid for specific time 3 The transaction timing constraints ket transactions to general-purpose computing systems that run audio and video applications. Increasingly computer systems even general-purpose systems are requiring real-time support so it s not surprising to hear more about real-time databases. Unfortunately there are many misconceptions about the real-time aspects of databases. Ironically the state of confusion that exists today about real-time databases parallels the confusion that existed a decade ago surrounding the differences between real-time and general-purpose We believe that a careful definition of real-time databases will help dispel these misconceptions and will encourage research efforts similar to those that have advanced real-time systems over this past decade. REAL-TIME DATABASES SOME DEFINITIONS We must first note that a system using real-time data such as sensor data does not in itself constitute a realtime database system. Because a real-time database is by definition a database system it has queries schemas transactions commit protocols concurrency control support and storage management. can be completion deadlines start times periodic invocations and so on. It is not necessary that every transaction have a timing constraint only that some do. In addition to transaction timing requirements data has time semantics as .