Towards using SCTP as a Data Transport Protocol for Data-Intensive Batch Jobs

Lucian Ghinea (1) , Mugurel Ionut Andreica (1) , Vlad Olaru (1) , Nicolae Tapus (1)

1 Parallel and Distributed Systems Laboratory [Bucarest]

In this paper we investigate the usage of the Stream Control Transmission Protocol (SCTP) as a data transport protocol for client-initiated data-intensive computations. The considered environment consists of job schedulers, storage nodes, computational nodes (workers) and clients. The clients submit jobs to the schedulers, which split them into multiple computation units and schedule these computations on the available workers. The data which needs to be processed by the computation units will be transferred from the storage nodes to the computational nodes using SCTP. Experimental evaluations considered images as the data to be processed and simple image processing operations as computations.

Mots clés: data-intensive computations; SCTP; SOAP; workers; image processing

Domaines: Réseaux et télécommunications [cs.NI]; Calcul parallèle, distribué et partagé [cs.DC]

Lucian Ghinea, Mugurel Ionut Andreica, Vlad Olaru, Nicolae Tapus. Towards using SCTP as a Data Transport Protocol for Data-Intensive Batch Jobs. Proceedings of the 19th International Conference on Control Systems and Computer Science (CSCS) (ISBN: 978-1-4673-6140-8), May 2013, Bucharest, Romania. pp.83-90, ⟨10.1109/CSCS.2013.29⟩. ⟨hal-00803423⟩