TY - GEN
T1 - On merits and viability of multi-cloud serverless
AU - Baarzi, Ataollah Fatahi
AU - Kesidis, George
AU - Joe-Wong, Carlee
AU - Shahrad, Mohammad
N1 - Funding Information:
Another line of research aims to build algorithms to dynamically utilize multiple cloud services or providers in order to minimize cost. Much of this work has focused on utilizing Amazon EC2’s spot and burstable offerings, which offer discounted services at lower availability [6, 37, 55]. These works generally attempt to minimize user costs while participating in multiple markets, with the hope that at least one market will have available resources that the user can utilize at any given time. Thus, their focus is on managing availability by exploiting the dynamics of the spot market. Other work [56] has examined the viability of a virtual provider that aggregates user jobs at multiple cloud providers in order to save costs. However, the temporal variation in serverless computing performance, as well as the relative complexity of its pricing policies, likely require the VSP to develop new scheduling algorithms for mapping functions to providers. Acknowledgments We thank David Wentzlaff, Yiying Zhang, and Samuel Ginzburg for their valuable feedback on this work. We also thank the anonymous reviewers and our shepherd, Dongsu Han, for helping us improve the paper. This work was supported in part by NSF CCF grant 2028929, NSF CNS grant 2122155, NSF CNS grant 1751075, NSERC grant RGPIN-2021-03714, and the AWS Cloud Credits for Research program.
Publisher Copyright:
© 2021 Association for Computing Machinery.
PY - 2021/11/1
Y1 - 2021/11/1
N2 - Serverless computing is a rapidly growing paradigm in the cloud industry that envisions functions as the computational building blocks of an application. Instead of forcing the application developer to provision cloud resources for their application, the cloud provider provisions the required resources for each function "under the hood."In this work, we envision virtual serverless providers (VSPs) to aggregate serverless offerings. In doing so, VSPs allow developers (and businesses) to get rid of vendor lock-in problems and exploit pricing and performance variation across providers by adaptively utilizing the best provider at each time, forcing the providers to compete to offer cheaper and superior services. We discuss the merits of a VSP and show that serverless systems are well-suited to cross-provider aggregation, compared to virtual machines. We propose a VSP system architecture and implement an initial version. Using experimental evaluations, our preliminary results show that a VSP can improve maximum sustained throughput by 1.2x to 4.2x, reduces SLO violations by 98.8%, and improves the total invocations' costs by 54%.
AB - Serverless computing is a rapidly growing paradigm in the cloud industry that envisions functions as the computational building blocks of an application. Instead of forcing the application developer to provision cloud resources for their application, the cloud provider provisions the required resources for each function "under the hood."In this work, we envision virtual serverless providers (VSPs) to aggregate serverless offerings. In doing so, VSPs allow developers (and businesses) to get rid of vendor lock-in problems and exploit pricing and performance variation across providers by adaptively utilizing the best provider at each time, forcing the providers to compete to offer cheaper and superior services. We discuss the merits of a VSP and show that serverless systems are well-suited to cross-provider aggregation, compared to virtual machines. We propose a VSP system architecture and implement an initial version. Using experimental evaluations, our preliminary results show that a VSP can improve maximum sustained throughput by 1.2x to 4.2x, reduces SLO violations by 98.8%, and improves the total invocations' costs by 54%.
UR - http://www.scopus.com/inward/record.url?scp=85119250537&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85119250537&partnerID=8YFLogxK
U2 - 10.1145/3472883.3487002
DO - 10.1145/3472883.3487002
M3 - Conference contribution
AN - SCOPUS:85119250537
T3 - SoCC 2021 - Proceedings of the 2021 ACM Symposium on Cloud Computing
SP - 600
EP - 608
BT - SoCC 2021 - Proceedings of the 2021 ACM Symposium on Cloud Computing
PB - Association for Computing Machinery, Inc
T2 - 12th Annual ACM Symposium on Cloud Computing, SoCC 2021
Y2 - 1 November 2021 through 4 November 2021
ER -