A High Performance Concurrency Control Protocol for Multi-Processor Transaction Processing Systems

  • Shiwei Wang
  • Ugo O. Gagliardi


In this paper we have proposed a new concurrency control algorithm that consists of two phases of execution for centralized multiprocessor-based transaction processing systems. The proposed algorithm integrates optimistic concurrency control with back-shifting with pre-claimed locking schemes into two phases. It guarantees transactions to commit in two executions if access invariance holds for the second run. It also can offer superior performance than the existing concurrency algorithms by reduce the probability of aborts in the first phase, minimize the probability of blocking due to validation and the possible second phase. Furthermore, it is deadlock-free.


Data Block Concurrency Control Validation Stage Host Processor Serialization Graph 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    Kung and Robinson, “On Optimistic Methods for Concurrency Control,” ACM TODS 6 (2), 1981, 213–226.CrossRefGoogle Scholar
  2. [2]
    Yu and Dias, “Performance Analysis of Optimistic Concurrency Control Schemes for Systems with Large Memory,” Proc. ACM SIGMETRICS Conference, 1989.Google Scholar
  3. [3]
    Franazek, Robinson, and Thomasian “Access Invariance and Its Use in High Contention Environments,” Proc International Conf. on Data Engineering, 1990, 47–55.Google Scholar
  4. [4]
    Yu and Dias, “Concurrency Control Using Locking with Deferred Blocking,” Proc. Inter. Conf. on Data Engineering, 1990, 30–36.Google Scholar
  5. [5]
    Robinson, “Design of Concurrency Controls for Transaction Processing Systems,” Ph.D. thesis, tech. report, CMU-CS-82–114, CMU, 1982.Google Scholar
  6. [6]
    Wang and Li, “The Precedence-agreement for Distributed Database Concurrency Control algorithms,” Proc. ACM Symp. Principles of Database Systems, 1987, 119–128.Google Scholar
  7. [7]
    Agrawal, Carey, and Ltvny, “Concurrency Control Performance Modeling: Alternatives and Implications,” ACM TODS 12 (4), 609–654.Google Scholar
  8. [8]
    Gray, Lorie, Putzolu, “Granularity of Locks in a Shared Data Base,” Proc. International Conf. on Very Large Data Bases, 1975, 428–451.Google Scholar
  9. [9]
    Reuter, “The Transaction Pipeline Processor,” Proc, International Workshop on High Performance Transaction Systems, 1985.Google Scholar
  10. [10]
    Franaszek and Robinson, “Limitations of Concrrency in Transaction Processing,’ ACM TODS. March, 1985, 1–28.Google Scholar
  11. [11]
    Holt, ‘Some deadlock Properties in Computer Systems.’ ACM Computing Surveys 4 (3), 1972, 179–196.MathSciNetCrossRefGoogle Scholar

Copyright information

© Springer-Verlag/Wien 1990

Authors and Affiliations

  • Shiwei Wang
    • 1
  • Ugo O. Gagliardi
    • 1
  1. 1.Aiken Computation Lab., Division of Applied SciencesHarvard UniversityCambridgeUSA

Personalised recommendations