Institutional Repository [SANDBOX]
Technical University of Crete
EN  |  EL

Search

Browse

My Space

Smilax: statistical machine learning autoscaler agent for Apache Flink

Giannakopoulos Panagiotis, Petrakis Evripidis

Full record


URI: http://purl.tuc.gr/dl/dias/EA294D54-5302-4702-96B3-788F482F12BE
Year 2021
Type of Item Conference Publication
License
Details
Bibliographic Citation P. Giannakopoulos, and E. G. M. Petrakis, “Smilax: statistical machine learning autoscaler agent for Apache Flink,” in Advanced Information Networking and Applications, vol. 226, Lecture Notes in Networks and Systems, L. Barolli, I. Woungang, T. Enokido, Eds., Cham, Switzerland: Springer Nature, 2021, pp. 433–444, doi: 10.1007/978-3-030-75075-6_35. https://doi.org/10.1007/978-3-030-75075-6_35
Appears in Collections

Summary

Smilax is a statistical machine learning autoscaler agent for applications running on Apache Flink. Smilax agent acts proactively by predicting the forthcoming workload in order to adjust the allocation of workers to the actual needs of an application ahead of time. During an online training phase, Smilax builds a model which maps the performance of the application to the minimum number of servers. During the work (optimal) phase, Smilax maintains the performance of the application within acceptable limits (i.e. defined in the form of SLAs) while minimizing the utilization of resources. The effectiveness of Smilax is assessed experimentally by running a data intensive fraud detection application.

Services

Statistics