TY - GEN
T1 - Move fast and meet deadlines
T2 - 18th USENIX Symposium on Networked Systems Design and Implementation, NSDI 2021
AU - Xu, Le
AU - Venkataraman, Shivaram
AU - Gupta, Indranil
AU - Mai, Luo
AU - Potharaju, Rahul
N1 - Funding Information:
This work was supported in part by the following grants: NSF IIS 1909577, NSF CNS 1908888, NSF CNS 1319527, NSF CNS 1838733, a Facebook faculty research award, the Helios project at Microsoft [77], and another generous gift from Microsoft. Shivaram Venktaraman is also supported by the Office of the Vice Chancellor for Research and Graduate Education with funding from theWisconsin Alumni Research Foundation. We are grateful to the Cosmos, Azure Data Lake, and PlayFab teams at Microsoft.
Funding Information:
We thank our shepherd Matei Zaharia and our anonymous referees for their reviews and help with improving the paper. We thank Kai Zeng for providing feedbacks for initial ideas. This work was supported in part by the following grants: NSF IIS 1909577, NSF CNS 1908888, NSF CNS 1319527, NSF CNS 1838733, a Facebook faculty research award, the Helios project at Microsoft [77], and another generous gift from Microsoft. Shivaram Venktaraman is also supported by the Office of the Vice Chancellor for Research and Graduate Education with funding from the Wisconsin Alumni Research Foundation. We are grateful to the Cosmos, Azure Data Lake, and PlayFab teams at Microsoft.
Publisher Copyright:
© 2021 by The USENIX Association.
PY - 2021
Y1 - 2021
N2 - Resource provisioning in multi-tenant stream processing systems faces the dual challenges of keeping resource utilization high (without over-provisioning), and ensuring performance isolation. In our common production use cases, where streaming workloads have to meet latency targets and avoid breaching service-level agreements, existing solutions are incapable of handling the wide variability of user needs. Our framework called Cameo uses fine-grained stream processing (inspired by actor computation models), and is able to provide high resource utilization while meeting latency targets. Cameo dynamically calculates and propagates priorities of events based on user latency targets and query semantics. Experiments on Microsoft Azure show that compared to state-of-the-art, the Cameo framework: i) reduces query latency by 2.7× in single tenant settings, ii) reduces query latency by 4.6× in multi-tenant scenarios, and iii) weathers transient spikes of workload.
AB - Resource provisioning in multi-tenant stream processing systems faces the dual challenges of keeping resource utilization high (without over-provisioning), and ensuring performance isolation. In our common production use cases, where streaming workloads have to meet latency targets and avoid breaching service-level agreements, existing solutions are incapable of handling the wide variability of user needs. Our framework called Cameo uses fine-grained stream processing (inspired by actor computation models), and is able to provide high resource utilization while meeting latency targets. Cameo dynamically calculates and propagates priorities of events based on user latency targets and query semantics. Experiments on Microsoft Azure show that compared to state-of-the-art, the Cameo framework: i) reduces query latency by 2.7× in single tenant settings, ii) reduces query latency by 4.6× in multi-tenant scenarios, and iii) weathers transient spikes of workload.
UR - http://www.scopus.com/inward/record.url?scp=85106168340&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85106168340&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85106168340
T3 - Proceedings of the 18th USENIX Symposium on Networked Systems Design and Implementation, NSDI 2021
SP - 389
EP - 405
BT - Proceedings of the 18th USENIX Symposium on Networked Systems Design and Implementation, NSDI 2021
PB - USENIX Association
Y2 - 12 April 2021 through 14 April 2021
ER -