BALLAST: Bandit-Assisted Learning for Latency-Aware Stable Timeouts in Raft
Fuente:
arXiv
Gespeichert in:
| 1. Verfasser: | |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866908731240873984 |
|---|---|
| author | Wang, Qizhi |
| author_facet | Wang, Qizhi |
| contents | Randomized election timeouts are a simple and effective liveness heuristic for Raft, but they become brittle under long-tail latency, jitter, and partition recovery, where repeated split votes can inflate unavailability. This paper presents BALLAST, a lightweight online adaptation mechanism that replaces static timeout heuristics with contextual bandits. BALLAST selects from a discrete set of timeout "arms" using efficient linear contextual bandits (LinUCB variants), and augments learning with safe exploration to cap risk during unstable periods. We evaluate BALLAST on a reproducible discrete-event simulation with long-tail delay, loss, correlated bursts, node heterogeneity, and partition/recovery turbulence. Across challenging WAN regimes, BALLAST substantially reduces recovery time and unwritable time compared to standard randomized timeouts and common heuristics, while remaining competitive on stable LAN/WAN settings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_21165 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | BALLAST: Bandit-Assisted Learning for Latency-Aware Stable Timeouts in Raft Wang, Qizhi Machine Learning Artificial Intelligence 68M14 (Primary) 68W15, 68T05 (Secondary) C.2.4; D.4.5; I.2.6 Randomized election timeouts are a simple and effective liveness heuristic for Raft, but they become brittle under long-tail latency, jitter, and partition recovery, where repeated split votes can inflate unavailability. This paper presents BALLAST, a lightweight online adaptation mechanism that replaces static timeout heuristics with contextual bandits. BALLAST selects from a discrete set of timeout "arms" using efficient linear contextual bandits (LinUCB variants), and augments learning with safe exploration to cap risk during unstable periods. We evaluate BALLAST on a reproducible discrete-event simulation with long-tail delay, loss, correlated bursts, node heterogeneity, and partition/recovery turbulence. Across challenging WAN regimes, BALLAST substantially reduces recovery time and unwritable time compared to standard randomized timeouts and common heuristics, while remaining competitive on stable LAN/WAN settings. |
| title | BALLAST: Bandit-Assisted Learning for Latency-Aware Stable Timeouts in Raft |
| topic | Machine Learning Artificial Intelligence 68M14 (Primary) 68W15, 68T05 (Secondary) C.2.4; D.4.5; I.2.6 |
| url | https://arxiv.org/abs/2512.21165 |