A study of two transaction-processing architectures for distributed real-time data base systems

A real-time data base system (RTDBS) is designed to provide timely response to the transactions of data-intensive applications. Processing a transaction in a distributed RTDBS environment presents the design choice of how to provide access to remote data referenced by the transaction. Satisfaction of the timing constraints of transactions should be the primary factor to be considered in scheduling accesses to remote data. In this article, we describe and analyze two different alternative approaches to this fundamental design decision. With the first alternative, transaction operations are executed at the sites where required data pages reside. The other alternative is based on transmitting data pages wherever they are needed. Although the latter approach is characterized by large message volumes carrying data pages, it is shown in our experiments to perform better than the other approach under most of the work loads and system configurations tested. The performance metric used in the evaluations is the fraction of transactions that satisfy their timing constraints.

[1]  Hector Garcia-Molina,et al.  Scheduling Real-Time Transactions with Disk Resident Data , 1989, VLDB.

[2]  Miron Livny,et al.  Distributed Concurrency Control Performance: A Study of Algorithms, Distribution, and Replication , 1988, VLDB.

[3]  Lawrence A. Rowe,et al.  Cache consistency and concurrency control in a client/server DBMS architecture , 1991, SIGMOD '91.

[4]  H. Garcia-Molina,et al.  Reliable distributed database management , 1987, Proceedings of the IEEE.

[5]  A RoweLawrence,et al.  Cache consistency and concurrency control in a client/server DBMS architecture , 1991 .

[6]  Miron Livny,et al.  On being optimistic about real-time constraints , 1990, PODS '90.

[7]  Vlad Ingar Wietrzyk,et al.  Real-Time Transaction Scheduling in Database Systems , 1996, DEXA.

[8]  Miron Livny,et al.  Global Memory Management in Client-Server Database Architectures , 1992, VLDB.

[9]  Herb Schwetman,et al.  CSIM: a C-based process-oriented simulation language , 1986, WSC '86.

[10]  Divyakant Agrawal,et al.  Using delayed commitment in locking protocols for real-time databases , 1992, SIGMOD '92.

[11]  M. Franklin,et al.  Global Memory Management in Client-Server DBMS Architectures , 1992 .

[12]  Donald F. Towsley,et al.  On using priority inheritance in real-time databases , 1991, [1991] Proceedings Twelfth Real-Time Systems Symposium.

[13]  Yi Lin,et al.  An integrated real-time locking protocol , 1992, [1992] Eighth International Conference on Data Engineering.

[14]  Hector Garcia-Molina,et al.  Scheduling real-time transactions: a performance evaluation , 1988, TODS.

[15]  Jaideep Srivastava,et al.  Enhancing real-time DBMS performance with multiversion data and priority based disk scheduling , 1991, [1991] Proceedings Twelfth Real-Time Systems Symposium.

[16]  Ophir Frieder Communications issues in data engineering: 'have bandwidth-will move data' , 1989, [1989] Proceedings. Fifth International Conference on Data Engineering.