CALM: A Self-Adaptive Orchestration Approach for QoS-Aware Routing in Small Language Model based Systems

TL;DR

CALM is a self-adaptive orchestration system that dynamically manages a fleet of Small Language Models to optimize QoS metrics like latency and energy consumption in real-time, addressing uncertainties in LLM-based systems.

Contribution

This paper introduces CALM, a novel MAPE-K based self-adaptive framework for orchestrating multiple SLMs to improve QoS in LLM-enabled systems.

Findings

Reduces latency by approximately 40%

Lowers energy consumption by 50%

Maintains task performance with dynamic SLM management

Abstract

AI-enabled systems are subjected to various types of runtime uncertainties, ranging from dynamic workloads, resource requirements, model drift, etc. These uncertainties have a big impact on the overall Quality of Service (QoS). This is particularly true in the case of Language Model (LM) enabled systems where the autoregressive nature of token generation introduces variability in latency, energy usage and response quality. These systems, powered by LLMs, are either resource-intensive (if run on-prem) or raise privacy/cost concerns (if leveraged using APIs). While deploying a Small Language Model (SLM) can be resource-efficient, it often falls short in addressing the diversity and scale of real-world requirements. To this, we argue that, rather than relying on any one SLM, leveraging a coordinated fleet of SLMs, each with specialized strengths can enable systems to dynamically adapt to shifting contexts and workload patterns. However, realizing the full potential of such an approach demands intelligent orchestration and continuous adaptation. To this end, we introduce CALM , a self-adaptive orchestration mechanism based on MAPE-K. Our approach continuously monitors user queries, analyzes the QoS metrics of the SLMs, identifies the optimal SLM to be used, routes the query to the identified SLM and further to enhance the effectiveness and efficiency, leverages caching and scheduling to decide the SLMs to be kept in memory. Our evaluation shows that CALM reduces latency by approximately 40% and energy consumption by 50%, while preserving domain-specific task performance when compared to single-LLM baselines.