ePrints@IIScePrints@IISc Home | About | Browse | Latest Additions | Advanced Search | Contact | Help

Sizing and Coordinated Scaling of Microservices

Agarwal, P and Lakshmi, J (2020) Sizing and Coordinated Scaling of Microservices. In: 9th IEEE International Conference on Cloud Computing in Emerging Markets, 04-07 Nov 2020, pp. 1-8.

[img] PDF
CCEM_01-08_2020.pdf - Published Version
Restricted to Registered users only

Download (546kB) | Request a copy
Official URL: https://doi.org/10.1109/CCEM50674.2020.00013


Microservice(MS) based architecture is recently drawing attention for its ability to offer ease of deployment, agile independent scaling, and fault-tolerance. This nature of architecting an application has been promoted with the idea of scaling. Containers enable faster MSes instantiation aiding agility in deployment and scaling. The ability of MS independent scaling can be exploited to find the right-size of the MS to scale with, where the size of the MS is with respect to the resources allocated to it. As a result, any sizing strategy used in scaling needs to be quick to aid in decision making. Apart from this, another interesting and important fact with MSes is that while they benefit from independent scaling, they still belong to a single application. Each functionality of the application is inter-related to fulfill a business goal. Also, the business logic is realized using different MSes, often creating a workflow. This introduces correlations among MSes for scaling. Intuitively, such correlations can provide early indications of workload variations which can be used during scaling to enhance application performance. This work focuses on designing scaling algorithm which explores such correlations to auto-scale the MS based applications. The paper also proposes a low computational complexity MS sizing algorithm that finds the right-size of the MSes, using heuristics based on workload variation trends observed from past usage history. The proposed sizing algorithm is compared with several published sizing strategies and the results indicate that the cost benefits achieved are comparable to other algorithms in the state-of-art, while being sufficiently faster than others. And, the coordinated scaling algorithm proposed in the paper provides significant benefits yielding to more than 50 reduction in SLA violations as compared to independent MS scaling approaches. © 2020 IEEE.

Item Type: Conference Paper
Publication: Proceedings - 2020 IEEE International Conference on Cloud Computing in Emerging Markets, CCEM 2020
Publisher: Institute of Electrical and Electronics Engineers Inc.
Additional Information: The copyright for this article belongs to Institute of Electrical and Electronics Engineers Inc.
Keywords: Cloud computing; Commerce; Decision making; Fault tolerance, Application performance; Business goals; Business logic; Low computational complexity; Scaling algorithm; Sizing algorithms; Usage history; Workload variation, Computational complexity
Department/Centre: Division of Interdisciplinary Sciences > Supercomputer Education & Research Centre
Date Deposited: 03 Dec 2021 08:31
Last Modified: 03 Dec 2021 08:31
URI: http://eprints.iisc.ac.in/id/eprint/70217

Actions (login required)

View Item View Item