Implementation of the Database Machine Direct

Boral, Haran; DeWitt, David J.; Friedland, Dina; Jarrell, N.F.; Wilkinson, W. Kevin

doi:10.1109/tse.1982.235882

“…This is a metric that accounts only for processing functionality, and does not describe the I/O implications of the different organizations of multiple DP systems. It also does not evaluate any of the other differences in database machine performance that are dependent on the number and organization of the processors, such as algorithms for performing aggregate functions [BORA80].…”

Section: Resultsmentioning

confidence: 99%

“…A complete, analytical exploration of -the differences in join algorithms for database machines is outside the range of this paper, and exists in [BORA80) • However, [BORA80] does not directly address the problem of disk-associated multiple processors, and we desire an approximation for the performance application.…”

Section: Bibliographic Searchmentioning

confidence: 99%

The effect of target applications on the design of database machines

Hawthorn

¹

1981

Proceedings of the 1981 ACM SIGMOD International Conference on Management of Data - SIGMOD '81

7

0

View full text Add to dashboard Cite

LBL-12412Specialized, single function processors can be built to be faster and cheaper than general purpose processors. Most database machines use such special purpose processors to manipulate data, with a general purpose managing processor to control the special purpose processors and perform utility functions.In this paper, the organization and use of these data manipulation processors is explored. Database machines are classified into single data manipulation processor systems, multiple disk-associated data manipulation processor systems, and multiple cache-associated processor systems.Examples of actual database machines are given for each category.Application types are classified into business, bibliographic search, and statistical analysis systems. A metric is developed to compare the performance of the categories of database machines with respect to the application types. The metric is the effective instruction rate, which is cornparable to the instruction rate (millions of instructions per second) for conventional computers.It is shown that the effective instruction rate is highly ~ensitive to the proportion of work performed in the database machine's data manipulation processors.Therefore, determining the work that must be performed in the machine's managing processor is found to be important to determining the performance of the machine.Database machine performance for each category of database machines is compared for each type of application. It is shown that single processor systems are best for business applications; that disk-based multiple processor systems are best for bibliographic search applications; and that hybrid systems are best for statistical analysis applications. Therefore, the design of the organization of data manipulation processors is shown to be application-dependent.

show abstract

“…A complete, analytical exploration of -the differences in join algorithms for database machines is outside the range of this paper, and exists in [BORA80) • However, [BORA80] does not directly address the problem of disk-associated multiple processors, and we desire an approximation for the performance application.…”

Section: Business Applicationsmentioning

confidence: 99%

The effect of target applications on the design of database machines

Hawthorn¹

1981

Proceedings of the 1981 ACM SIGMOD International Conference on Management of Data - SIGMOD '81

View full text Add to dashboard Cite

LBL-12412Specialized, single function processors can be built to be faster and cheaper than general purpose processors. Most database machines use such special purpose processors to manipulate data, with a general purpose managing processor to control the special purpose processors and perform utility functions.In this paper, the organization and use of these data manipulation processors is explored. Database machines are classified into single data manipulation processor systems, multiple disk-associated data manipulation processor systems, and multiple cache-associated processor systems.Examples of actual database machines are given for each category.Application types are classified into business, bibliographic search, and statistical analysis systems. A metric is developed to compare the performance of the categories of database machines with respect to the application types. The metric is the effective instruction rate, which is cornparable to the instruction rate (millions of instructions per second) for conventional computers.It is shown that the effective instruction rate is highly ~ensitive to the proportion of work performed in the database machine's data manipulation processors.Therefore, determining the work that must be performed in the machine's managing processor is found to be important to determining the performance of the machine.Database machine performance for each category of database machines is compared for each type of application. It is shown that single processor systems are best for business applications; that disk-based multiple processor systems are best for bibliographic search applications; and that hybrid systems are best for statistical analysis applications. Therefore, the design of the organization of data manipulation processors is shown to be application-dependent.

show abstract

“…Recently, there has been much effort to improve the performance of database machines by achieving more parallelism in their architectures such as MICRONET [Su, 1978], EDC [Uemura, 1980], MDBS [Hsiao, 1983], DBC/1012~eradata, 1984, GAMMA [DeWkt, 1986], REPT [Shultq 1984], HYPERTREE [Despain, 1978;Sequine, 1978], DIRECT [Boral, 1982], GRACE [Kltsuregawa, 1985], DBMAC [Missikoff, 1981;Missikoff, lW, Missikoff, 19@ Cesaritti, 1983], and SM3 [Bar% 1986]. The major advantages of using a multicomputer architecture are 1) component reliability, 2) component availability, 3) cost-effectiveness, and 4) expandability.…”

Section: Introductionmentioning

confidence: 99%

Performance comparison of join on hypercube and mesh

Jhang

¹

1992

Proceedings of the 1992 ACM Annual Conference on Communications - CSC '92

View full text Add to dashboard Cite

In this paper, we compare the performance of two join algorithms on a cube and on a multidimensional mesh architectures. The algorithms are nested-loop and sortmerge joins. The study confirms that a higher connectivity architecture is faster and has a wider performance bandwidth. The sort-merge join algorithm is better for a small number of nodes, but it becomes worse than the nested-loop join algorithm as the number of nodes gets larger and the join selectivity gets large, especially for the simpler mesh architecture. The study also shows that the 3-dimensional mesh architecture has performanm comparable to the more complicated cube architecture for the medium number of nodes in a system.

show abstract

Implementation of the Database Machine Direct

Cited by 28 publications

References 10 publications

The effect of target applications on the design of database machines

The effect of target applications on the design of database machines

The effect of target applications on the design of database machines

Performance comparison of join on hypercube and mesh

Contact Info

Product

Resources

About