M. Sullivan. System Support for Software Fault Tolerance in Highly Available Database Management Systems. PhD thesis, University of California, Berkeley, January 1993. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Fault Resilience In Main-Memory Databases: Handling Process.. - Bohannon (1999) (Correct)
....was reviewed in the Introduction, and showed that the percentage of errors attributable to software rose from 33 to 62 between 1985 and 1989. In his thesis, Sullivan reviews field errors recorded by IBM in internal databases for the MVS operating system and two database products, IMS and DB2 [74]. His study categorized errors as control errors, addressing errors, and data errors. Control errors affect variables that represent the current state of the program. Data errors cause the result of the computation to be wrong, but do not cause the system to become mistaken about its own state. ....
....cost of RAM ensures that commercial databases will approach the limit case by caching more and more memory for longer periods of time. Comparing the two studies illustrates how difficult it is to predict errors scientifically. In particular, addressing errors, found to be 25 30 of the errors in [74] contributed to none of the errors found by Ng and Chen in [59] Thus, even though Ng and Chen weighted their results by Sullivan s numbers, this category did not weigh in their results at all. Almost certainly this results from the difference between a fault injection scenario and field errors. ....
M. Sullivan. System Support for Software Fault Tolerance in Highly Available Database Management Systems. PhD thesis, University of California, Berkeley, January 1993.
Detection and Recovery Techniques for Database Corruption - Bohannon, Rastogi.. (Correct)
....Domain Sizes . 30 3 Cost of Corruption Protection . 31 1 Introduction As hardware gets more reliable, software errors are often the greatest threat to database system availability [Sul93, Gra90] even in standard database systems where database data is protected by process boundaries from errors in application programs. Increasingly, however, for extensibility and performance, special purpose application code is being integrated with database system code. Extensible databases ....
....performance needs of these applications. Thus, database availability can be affected not only by software errors in the DBMS, but also by errors in application programs. One class of software error which has been shown to have a significant impact on DBMS availability is the addressing error [Sul93] This class of error includes copy overruns and wild writes through uninitialized pointers. Software fault tolerance techniques (see, for example, MT77, Ran77] attempt to mitigate the damage done when software errors occur in a production environment. One approach to avoiding addressing ....
Mark Sullivan. System support for software fault tolerance in highly available database management systems. Technical Report ERL-93-5, University of California, Berkeley, January 1993.
Using Codewords to Protect Database Data from a.. - Bohannon.. (1999) (1 citation) (Correct)
....low cost logging of transaction reads to trace the corruption, and may also prove useful when resolving problems caused by incorrect data entry and other logical errors. 1. Introduction As hardware gets more reliable, software errors are often the greatest threat to database system availability [20, 7], even in standard systems where database data is protected by process boundaries from errors in application programs. Increasingly, however, for extensibility and performance, The work of S. Sudarshan was performed in part while at Bell Labs. special purpose application code is being integrated ....
....the performance needs of these applications. Thus, database availability can be affected not only by software errors in the DBMS, but also by errors in application programs. One class of software error which has been shown to have a significant impact on DBMS availability is the addressing error [20]. This class of error includes copy overruns and wild writes through uninitialized pointers. Software fault tolerance techniques (see, for example, 15, 18] attempt to mitigate the damage done when software errors occur in a production environment. One approach to avoiding addressing errors is ....
M. Sullivan. System Support for Software Fault Tolerance in Highly Available Database Management Systems. PhD thesis, University of California, Berkeley, Jan. 1993.
Integrating Reliable Memory in Databases - Ng, Chen (1998) (6 citations) (Correct)
.... be written to disk if memory is large enough to contain the undo records for all transactions in progress [Agrawal89] In addition, storing the database buffer cache in reliable memory allows the system to begin operation after a crash with the contents present prior to the crash (a warm cache) [Sullivan93, Elhardt84, Bhide93]. Storing the database or the log in reliable memory is a more powerful technique than group commit (an optimization commonly used to improve transaction throughput) DeWitt84] Systems that perform group commit wait for a number of committing transactions to accumulate, then synchronously write ....
....to be precisely modeled, so we distribute their weights evenly across other categories. Our main conclusion is that mapping reliable memory directly into the database address space has only a small effect on the overall reliability of the system. This is consistent with the estimates given in [Sullivan91a, Sullivan93]. There are several factors that minimize the reliability impact of persistent buffer caches. First, most stores in Postgres are not to the buffer cache. Using the ATOM program analysis tool [Srivastava94] we found that only 2 3 of stores executed during a run were to the buffer cache. Second, ....
[Article contains additional citation context not shown here]
Sullivan MP (1993) System Support for Software Fault Tolerance in Highly Available Database Management Systems. PhD thesis, University of California at Berkeley, Calif.
Parallel Query Processing Using Shared Memory Multiprocessors and.. - Hong (1992) (12 citations) (Correct)
....implications of running XPRS on other disk array configurations will be discussed in Chapter 5. The goals of XPRS are high performance and high availability. This thesis only deals with the high performance aspects of XPRS. Discussions on the high availability aspects of XPRS can be found in [53]. XPRS is designed to achieve high performance for both transaction processing and complex ad hoc queries through parallelism within each individual query as well as between different queries. 18 Parallel Query Execution in Volcano The Volcano query processing system is also a shared everything ....
Mark Sullivan. System Support for Software Fault Tolerance in Highly Available Database Management Systems. PhD thesis, University of California at Berkeley, August 1992.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC