[Survey] Towards Efficient Large Language Model Serving: A Survey on System-Aware KV Cache Optimization
