[Router] Add bucket-aware policy domains and native cache indexing (#38108)

Signed-off-by: Vincent Gao <vincentbo@linux.alibaba.com>
Co-authored-by: inkcherry <mingzhi.liu@amd.com>
Co-authored-by: yangbodong22011 <13137470+yangbodong22011@users.noreply.github.com>
This commit is contained in:
Vincent Gao
2026-09-06 19:47:51 +08:00
committed by GitHub
co-authored by inkcherry yangbodong22011
parent a176ba2f7b
commit 5bebe7a033
76 changed files with 5842 additions and 3639 deletions
@@ -13,7 +13,7 @@ use sgl_router::config::{
pub const MODEL: &str = "deepseek-v4-tiny";
/// A single-model `cache_aware_zmq` router. Discovery is a placeholder because
/// A single-model native `cache_aware` router. Discovery is a placeholder because
/// every caller installs its own `WorkerRegistry`.
pub fn config() -> Config {
Config {
@@ -25,11 +25,13 @@ pub fn config() -> Config {
model: ModelConfig {
id: MODEL.into(),
tokenizer_path: "tests/fixtures/tiny_tokenizer.json".into(),
policy: PolicyKind::CacheAwareZmq,
policy: PolicyKind::CacheAware,
decode_policy: Default::default(),
bucket_config: None,
circuit_breaker: None,
cache_aware: Some(CacheAwareConfig::default()),
sticky: None,
affinity: None,
sticky: None,
fused: None,
eligibility: None,
},