Discoveries

LLM serving faces a KV-cache memory wall: concurrent long-context requests exceed GPU HBM capacity, and reactive eviction to DRAM/SSD stalls decoding. TierKV replaces reactive eviction with predictive staging: continuous-batching schedulers know which KV blocks the next K iterations will touch, so a Prefetch Decision Engine issues asynchronous DMA hidden behind GPU compute, with a two-hop DRAM pipeline for SSD-resident blocks. Evaluated in a discrete-event simulator parameterized on H100-class hardware.

Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI. It features a KVCache-centric disaggregated architecture that separates the prefill and decoding clusters. It also leverages the underutilized CPU, DRAM, and SSD resources of the GPU cluster to implement a disaggregated cache of KVCache. The core of Mooncake is its KVCache-centric scheduler, which balances maximizing overall effective throughput while meeting latency-related Service Level Objectives (SLOs). Unlike traditional studies that assume all requests will be processed, Mooncake faces challenges due to highly overloaded scenarios. To mitigate these, we developed a prediction-based early rejection policy. Experiments show that Mooncake excels in long-context scenarios. Compared to the baseline method, Mooncake can achieve up to a 525% increase in throughput in certain simulated scenarios while adhering to SLOs. Under real workloads, Mooncake's innovative architecture enables Kimi to handle 75% more requests.