Serving Systems

Prompt caching, gateways and routing, token accounting, and multi-tenant isolation.

13concepts
195flashcards
108minutes of reading
  1. 01 Admission Control and Load Shedding Why an overloaded LLM server degrades for everyone at once, how token-aware admission control keeps latency SLOs intact by rejecting work early, and what graceful degradation looks like when you would rather not say no. advanced 8m 18 cards
  2. 02 Disaggregated Prefill and Decode Serving Why prefill and decode want opposite hardware and parallelism, how splitting them across separate GPU pools raises goodput, and what the KV cache transfer costs. advanced 9m 15 cards
  3. 03 Multi-Tenant Serving and Isolation Serving many tenants from one model is cheap and easy; giving each tenant their own fine-tune is expensive and hard. S-LoRA and per-request LoRA serving collapse the trade-off, but only for tenants who can share a base model. advanced 10m 7 cards
  4. 04 Prefix-Aware Routing and KV Cache Reuse Why load-balancing LLM requests round-robin throws away computed KV cache, and how routing on prompt prefix turns a fleet's caches into a shared asset. advanced 8m 18 cards
  5. 05 Queueing Theory for LLM Serving Little's Law fixes the maximum request rate a batched LLM server can sustain at a given latency, and the heavy-tailed distribution of output lengths explains why queues form long before the GPU is saturated. advanced 8m 24 cards
  6. 06 Request Scheduling and Preemption Shortest-job-first minimises average waiting time and requires knowing job length, which an LLM server cannot know; the workarounds are predicting the rank of output lengths or preempting at token granularity. advanced 8m 24 cards