Publikation
ScaleEvict: Altruistic Eviction for RDMA-Enabled Distributed Storage Engines
Till Steinert; Muhammad El-Hindi; Tobias Ziegler; Viktor Leis; Carsten Binnig
In: Proceedings of the 22nd International Workshop on Data Management on New Hardware, DaMoN 2026, Bengaluru, India, 31 May 2026- 5 June 2026. International Workshop on Data Management on New Hardware (DaMoN), Pages 2:1-2:8, ACM, 2026.
Zusammenfassung
Modern hardware and economic trends are driving the adoption
of distributed storage engines that expose a transparent, shared-
cache abstraction: any node can access both the cluster’s aggregate
DRAM and its NVMe storage over a fast (RDMA) network. To sus-
tain performance under changing workloads, these systems must
continuously evict and re-cache pages at very high rates. How-
ever, most designs still rely on node-local eviction algorithms such
as LRU, which waste aggregate DRAM by retaining redundant
page copies. We propose ScaleEvict, an altruistic eviction strategy
implemented in the state-of-the-art ScaleStore engine. ScaleEvict
efficiently coordinates eviction decisions across nodes to reduce
redundant replication and prioritize globally valuable pages. ScaleE-
vict matches ScaleStore’s throughput using only two-thirds of the
DRAM. Alternatively, at equal DRAM capacity, ScaleEvict improves
throughput by up to 2× while reducing 99th percentile latency by
up to 3×.