Service-oriented applications do not fully benefit from standard APIs yet, and many applications fail to use interchangeably all the services that implement a standard service API. This paper presents an approach to develop adaptation strategies that improve service interchangeability for service-oriented applications based on standard APIs. In our approach, an adaptation strategy consists of sets of parametric adaptation plans (called test-and-adapt plans), which execute test cases to reveal the occurrence of interchangeability problems, and activate runtime adaptors according to the test results. Throughout the paper, we formalize the structure of the parametric test-and-adapt plans and of their execution semantics, present an algorithm for identifying correct execution orders through sets of test-and-adapt plans, provide empirical evidence of the occurrence of interchangeability problems for sample applications and services, and discuss the effectiveness of the approach in terms of avoided failures, runtime overheads and development costs.
Denaro, G., Pezze', M., Tosi, D. (2013). Test-and-Adapt: An Approach for Improving Service Interchangeability. ACM TRANSACTIONS ON SOFTWARE ENGINEERING AND METHODOLOGY, 22(4), 1-42 [10.1145/2522920.2522921].
Test-and-Adapt: An Approach for Improving Service Interchangeability
DENARO, GIOVANNI;PEZZE', MAURO;TOSI, DAVIDE
2013
Abstract
Service-oriented applications do not fully benefit from standard APIs yet, and many applications fail to use interchangeably all the services that implement a standard service API. This paper presents an approach to develop adaptation strategies that improve service interchangeability for service-oriented applications based on standard APIs. In our approach, an adaptation strategy consists of sets of parametric adaptation plans (called test-and-adapt plans), which execute test cases to reveal the occurrence of interchangeability problems, and activate runtime adaptors according to the test results. Throughout the paper, we formalize the structure of the parametric test-and-adapt plans and of their execution semantics, present an algorithm for identifying correct execution orders through sets of test-and-adapt plans, provide empirical evidence of the occurrence of interchangeability problems for sample applications and services, and discuss the effectiveness of the approach in terms of avoided failures, runtime overheads and development costs.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.