Adding Envoy Gateway + GIE support

[Xunzhuo]

1. Robust Traffic Management
2. Dynamic Configuration and filters
3. LLM Aware Routing

[BrianPark314]

hey @Xunzhuo , can you split these issue into separate issues? also could you be more specific on the requirements and why they are needed?

[Xunzhuo]

Currently, we impl a python router for llm inference routing, it is lightweight, but lack of production-ready abilitiy related to LLM Traffic, for example:

1. Servcie Government
2. Security Management
3. Traffic Management
4. Performance & Stability

Introducing an AI GW, can implement what we already did in router like prefix-cache aware, also guarantee the points above. We need a stable, high-performance, observability, extensible proxy as a dataplane.

@BrianPark314 @YuhanLiu11