Real-Time Transaction Scheduling: A Framework for SynthesizingStatic and Dynamic Factors

  • Authors:

  • Sharma Chakravarthy;Dong-Kweon Hong;Theodore Johnson

  • Affiliations:

  • Database Systems Research and Development Center, Computer and Information Science and Engineering Department, University of Florida, Gainesville, FL 32611;Database Systems Research and Development Center, Computer and Information Science and Engineering Department, University of Florida, Gainesville, FL 32611;Database Systems Research and Development Center, Computer and Information Science and Engineering Department, University of Florida, Gainesville, FL 32611

  • Venue:
  • Real-Time Systems
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

Real-time databases are poised to be an important componentof complex embedded real-time systems. In real-time databases(as opposed to real-time systems), transactions must satisfythe ACID properties in addition to satisfying the timing constraintsspecified for each transaction (or task). Although several approacheshave been proposed to combine real-time scheduling and databaseconcurrency control methods, to the best of our knowledge, noneof them provide a framework for taking into account the dynamiccost associated with aborts, rollbacks, and restarts of transactions. In this paper, we propose a framework in which both static anddynamic costs of transactions can be taken into account. Specifically,we present: i) a method for pre-analyzing transactions basedon the notion of branch-points for data accessed up to a branchpoint and predicting expected data access to be incurred forcompleting the transaction, ii) a formulation of cost that includesstatic and dynamic factors for prioritizing transactions, iii)a scheduling algorithm which uses the above two, and iv) simulationof the algorithm for several operating conditions and workload. Our dynamic priority assignment policy (termed the cost consciousapproach or CCA) adapts well to fluctuations in the system loadwithout causing excessive numbers of transaction restarts. Oursimulations indicate that i) CCA performs better than the EDF-HPalgorithm for both soft and firm deadlines, ii) CCA is more fairthan EDF-HP, iii) CCA is better than EDF-CR for soft deadline,even though CCA requires and uses less information, and iv) CCAis especially good for disk-resident data.