L7FP synthesizes eBPF data planes to enforce the majority of L7 policies in the kernel for service meshes, delivering up to 6x lower median latency and 3x higher throughput than state-of-the-art proxies while falling back for unsupported policies.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
verdicts
UNVERDICTED 2representative citing papers
BatchLLM achieves 1.3x-10.8x higher throughput than vLLM and SGLang for batched LLM inference with prefix sharing via global prefix identification, decoding-first reordering, and memory-centric token batching.
citing papers explorer
-
Offloading L7 Policies to the Kernel
L7FP synthesizes eBPF data planes to enforce the majority of L7 policies in the kernel for service meshes, delivering up to 6x lower median latency and 3x higher throughput than state-of-the-art proxies while falling back for unsupported policies.
-
BatchLLM: Optimizing Large Batched LLM Inference with Global Prefix Sharing and Throughput-oriented Token Batching
BatchLLM achieves 1.3x-10.8x higher throughput than vLLM and SGLang for batched LLM inference with prefix sharing via global prefix identification, decoding-first reordering, and memory-centric token batching.