A Performance Analyzer for a Public Cloud's ML-Augmented VM Allocator
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914535984594944 |
|---|---|
| author | Bostandoost, Roozbeh Namyar, Pooria Kakarla, Siva Kesava Reddy Beckett, Ryan Segarra, Santiago Cortez, Eli Mallick, Ankur Hsieh, Kevin Fonseca, Rodrigo Hajiesmaili, Mohammad Arzani, Behnaz |
| author_facet | Bostandoost, Roozbeh Namyar, Pooria Kakarla, Siva Kesava Reddy Beckett, Ryan Segarra, Santiago Cortez, Eli Mallick, Ankur Hsieh, Kevin Fonseca, Rodrigo Hajiesmaili, Mohammad Arzani, Behnaz |
| contents | Cloud operators increasingly deploy multiple ML models in their VM allocation pipelines. In such settings, individually benign predictions can shift and compound, severely degrading performance. In a cloud provider's VM placement pipeline, CPU, memory, and lifetime prediction models jointly determine server count, live migration frequency, and network utilization; yet no existing approach can systematically stress-test how these models adversely interact. Deterministic adversarial analyzers cannot capture probabilistic ML behavior, so operators miss failures that arise only from correlated distributional shifts across models
In SANJESH, we formulate a bi-level optimization that captures how the ML models behave statistically and uncovers how they adversely interact. The outer level searches over what predictions the ML models could produce under distributional uncertainty to find adversarial conditions; the inner level evaluates how the VM allocator behaves given those predictions. When we applied it to the operator's production traces, SANJESH uncovered scenarios that cause $4\times$ worse performance than the operators' evaluator detected. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_07750 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Performance Analyzer for a Public Cloud's ML-Augmented VM Allocator Bostandoost, Roozbeh Namyar, Pooria Kakarla, Siva Kesava Reddy Beckett, Ryan Segarra, Santiago Cortez, Eli Mallick, Ankur Hsieh, Kevin Fonseca, Rodrigo Hajiesmaili, Mohammad Arzani, Behnaz Distributed, Parallel, and Cluster Computing Cloud operators increasingly deploy multiple ML models in their VM allocation pipelines. In such settings, individually benign predictions can shift and compound, severely degrading performance. In a cloud provider's VM placement pipeline, CPU, memory, and lifetime prediction models jointly determine server count, live migration frequency, and network utilization; yet no existing approach can systematically stress-test how these models adversely interact. Deterministic adversarial analyzers cannot capture probabilistic ML behavior, so operators miss failures that arise only from correlated distributional shifts across models In SANJESH, we formulate a bi-level optimization that captures how the ML models behave statistically and uncovers how they adversely interact. The outer level searches over what predictions the ML models could produce under distributional uncertainty to find adversarial conditions; the inner level evaluates how the VM allocator behaves given those predictions. When we applied it to the operator's production traces, SANJESH uncovered scenarios that cause $4\times$ worse performance than the operators' evaluator detected. |
| title | A Performance Analyzer for a Public Cloud's ML-Augmented VM Allocator |
| topic | Distributed, Parallel, and Cluster Computing |
| url | https://arxiv.org/abs/2512.07750 |