Multi-queue service for task scheduling based on data availability

• Autorzy: Rybiński Kamil , Śmiałek Michał

Identyfikatory

DOI

10.15439/2023F8957

• Języki publikacji: EN

Abstrakty

EN

Large-scale computation (LSC) systems are often performed in distributed environments where message passing is the key to orchestrating computations. In this paper we present a new message queue concept developed within the context of an LSC system (BalticLSC). The concept consists in proposing a multi-queue, where queues are grouped into families. A queue family can be used to distribute messages of the same kind to multiple computation modules distributed between various nodes. Such message families can be synchronised to implement a mechanism for initiating computation jobs based on multiple data inputs. Moreover, the proposed multi-queue has built-in mechanisms for controlling message sequences in applications where complex data set splitting is necessary. The presented multi-queue concept was implemented and applied with success in a working LSC system.

• Wydawca: Polskie Towarzystwo Informatyczne
• Czasopismo: Annals of Computer Science and Information Systems
• Rocznik: 2023
• Tom: Vol. 35
• Strony: 699--709
• Opis fizyczny: Bibliogr. 25 poz.

Twórcy

autor

Rybiński Kamil

autor

Śmiałek Michał

Uwagi

1. Thematic Tracks Regular Papers

2. Opracowanie rekordu ze środków MEiN, umowa nr SONP/SP/546092/2022 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2024).