Enabling Resource Access Visibility for Automated Enterprise Services

Abstract

Organizations deliver on their mandates by executing a variety of services. Over the past few decades, service automation software systems, such as SAP and PeopleSoft, have enabled the automation of services. While much attention in the literature and in industry has been devoted to the implementation and functional correctness of automated services, little focus has been granted to ensuring responsiveness for services. As service automation platforms host larger and larger numbers of services, and services execute with greater and greater levels of concurrency, fault resolution becomes an important issue in ensuring expected responsiveness levels. In particular, two factors impact fault resolution in service automation platforms. First, each executing service requires access to specific data and system resources to complete its processing. As greater numbers of services execute concurrently, there is increasing contention for these data and system resources, leading to greater numbers of faults and SLA violations in service execution. Second, the black-box nature of service automation platforms provides little visibility into the nature of resource contention that caused a fault or SLA violation. This lack of visibility makes fault resolution difficult, and in many cases impossible, because it is difficult to trace the root cause of the problem. In this paper, the authors address the problem of system-level resource visibility for services through the design and development of a system capable of mapping abstract service workflows to their data and system impacts to enable resource visibility. The authors' system has been tested and demonstrated effective, as we demonstrate in a case study setting.