Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2212.05155 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913133032898560 |
|---|---|
| author | Ding, Yi Gao, Aijia Ryden, Thibaud Sedlak, Michal Ewaisha, Essam Marnat, Igor Hoffmann, Henry |
| author_facet | Ding, Yi Gao, Aijia Ryden, Thibaud Sedlak, Michal Ewaisha, Essam Marnat, Igor Hoffmann, Henry |
| contents | Software upgrades are critical to maintaining server reliability in datacenters. While job duration prediction and scheduling have been extensively studied, the unique challenges posed by software upgrades remain largely under-explored. This paper presents the first in-depth investigation into software upgrade scheduling at datacenter scale. We begin by characterizing various types of upgrades and then frame the scheduling task as a constrained optimization problem. To address this problem, we introduce Acela, a cost-aware duration prediction framework designed to improve upgrade scheduling efficiency and throughput while meeting service-level objectives (SLOs). Acela accounts for asymmetric misprediction costs, strategically selects the best predictive models, and mitigates straggler-induced overestimations. Evaluations on Meta's production datacenter systems demonstrate that Acela significantly increases efficiency of the existing upgrade scheduler by improving upgrade window utilization by 1.25X, increasing the number of scheduled and completed upgrades by 33% and 41%, and reducing cancellation rates by 2.4X. The code and data sets will be released after paper acceptance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2212_05155 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Cost-aware Duration Prediction for Software Upgrades in Datacenters Ding, Yi Gao, Aijia Ryden, Thibaud Sedlak, Michal Ewaisha, Essam Marnat, Igor Hoffmann, Henry Distributed, Parallel, and Cluster Computing Machine Learning Software upgrades are critical to maintaining server reliability in datacenters. While job duration prediction and scheduling have been extensively studied, the unique challenges posed by software upgrades remain largely under-explored. This paper presents the first in-depth investigation into software upgrade scheduling at datacenter scale. We begin by characterizing various types of upgrades and then frame the scheduling task as a constrained optimization problem. To address this problem, we introduce Acela, a cost-aware duration prediction framework designed to improve upgrade scheduling efficiency and throughput while meeting service-level objectives (SLOs). Acela accounts for asymmetric misprediction costs, strategically selects the best predictive models, and mitigates straggler-induced overestimations. Evaluations on Meta's production datacenter systems demonstrate that Acela significantly increases efficiency of the existing upgrade scheduler by improving upgrade window utilization by 1.25X, increasing the number of scheduled and completed upgrades by 33% and 41%, and reducing cancellation rates by 2.4X. The code and data sets will be released after paper acceptance. |
| title | Cost-aware Duration Prediction for Software Upgrades in Datacenters |
| topic | Distributed, Parallel, and Cluster Computing Machine Learning |
| url | https://arxiv.org/abs/2212.05155 |