Szczegóły publikacji
Opis bibliograficzny
Adaptation of workflow application scheduling algorithm to serverless infrastructure / Maciej PAWLIK, Paweł Banach, Maciej MALAWSKI // W: Euro-Par 2019: parallel processing workshops : Euro-Par 2019 international workshops : Göttingen, Germany, August 26–30, 2019 : revised selected papers / eds. Ulrich Schwardmann, [et al.]. — Cham : Springer Nature Switzerland AG, cop. 2020. — (Lecture Notes in Computer Science ; ISSN 0302-9743 ; LNCS 11997. Theoretical Computer Science and General Issues ; ISSN 0302-9743). — ISBN: 978-3-030-48339-5; e-ISBN: 978-3-030-48340-1. — S. 345–356. — Bibliogr. s. 355–356, Abstr. — Publikacja dostępna online od: 2020-05-29
Autorzy (3)
Słowa kluczowe
Dane bibliometryczne
ID BaDAP | 129056 |
---|---|
Data dodania do BaDAP | 2020-06-24 |
Tekst źródłowy | URL |
DOI | 10.1007/978-3-030-48340-1_27 |
Rok publikacji | 2020 |
Typ publikacji | materiały konferencyjne (aut.) |
Otwarty dostęp | |
Wydawca | Springer |
Konferencja | International Conference on Parallel and Distributed Computing |
Czasopisma/serie | Theoretical Computer Science and General Issues, Lecture Notes in Computer Science |
Abstract
Function-as-a-Service is a novel type of cloud service used for creating distributed applications and utilizing computing resources. Application developer supplies source code of cloud functions, which are small applications or application components, while the service provider is responsible for provisioning the infrastructure, scaling and exposing a REST style API. This environment seems to be adequate for running scientific workflows, which in recent years, have become an established paradigm for implementing and preserving complex scientific processes. In this paper, we present work done on adaptation of a scheduling algorithm to FaaS infrastructure. The result of this work is a static heuristic capable of planning workflow execution based on defined function pricing, deadline and budget. The SDBCS algorithm is designed to determine the quality of assignment of particular task to specific function configuration. Each task is analyzed for execution time and cost characteristics, while keeping track of parameters of complete workflow execution. The algorithm is validated through means of experiment with a set of synthetic workflows and a real life infrastructure case study performed on AWS Lambda. The results confirm the utility of the algorithm and lead us to propose areas of further study, which include more detailed analysis of infrastructure features affecting scheduling.