Organizing long-running activities with triggers and transactions
Umeshwar Dayal, Meichun Hsu, Rivka Ladin
Abstract
Open-access reader
Umeshwar Dayal, Meichun Hsu, Rivka Ladin
Abstract
Open-access reader
This paper addresses the problem of organising and controlling activities that involve multiple steps of processing and that typically are of long duration. We explore the use of triggers and transactions to specify and organize such long-running activities. Triggers offer data- or event-driven specification of control flow, and thus provide a flexible and modular framework with which the control structures of the activities can be extended or modified. We describe a model based on event-condition-action rules and coupling modes. The execution of these rules is governed by an extended nested transaction model. Through a detailed example, we illustrate the utility of the various features of the model for chaining related steps without sacrificing concurrency, for enforcing integrity constraints, and for providing flexible failure and exception handling.
OpenAlex reports 290 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
This paper addresses the problem of organising and controlling activities that involve multiple steps of processing and that typically are of long duration. We explore the use of triggers and transactions to specify and organize such long-running activities. Triggers offer data- or event-driven specification of control flow, and thus provide a flexible and modular framework with which the control structures of the activities can be extended or modified. We describe a model based on event-condition-action rules and coupling modes. The execution of these rules is governed by an extended nested transaction model. Through a detailed example, we illustrate the utility of the various features of the model for chaining related steps without sacrificing concurrency, for enforcing integrity constraints, and for providing flexible failure and exception handling.
Key concepts: Chaining, Computer science, Distributed computing, Event (particle physics), Concurrency, Modular design, Control flow, Database transaction