Replacing an SSD KV-offload tier with High-Bandwidth Flash in an SSD-style LLM serving stack raises average end-to-end latency 2 to 5.5 times and cuts SLO goodput, because transient KV is write-heavy and off the critical path.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AR 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
HBF Sucks! A Full-Stack Characterization of High-Bandwidth Flash for KV-Centric LLM Serving
Replacing an SSD KV-offload tier with High-Bandwidth Flash in an SSD-style LLM serving stack raises average end-to-end latency 2 to 5.5 times and cuts SLO goodput, because transient KV is write-heavy and off the critical path.