Aceso: Carbon-Aware and Cost-Effective Microservice Placement for Small and Medium-sized Enterprises

A team of researchers from institutions including Georgia Christofidi, Francisco Álvarez-Terribas, and Thaleia Dimitra Doudali has developed Aceso, an adaptive carbon- and efficiency-aware placement system specifically designed for small and medium-sized enterprises (SMEs) running microservice applications. Unlike existing solutions that target global-scale infrastructure or batch workloads, Aceso addresses the reality that most SMEs operate within geographically constrained regions with limited infrastructure options. The system dynamically places microservices using a scalable optimization strategy that incorporates real-time carbon intensity data, electricity costs, and latency requirements.

Aceso's key innovation lies in its insight-based search space pruning techniques, which allow it to quickly adapt to changing workloads and environmental conditions without overwhelming computational resources. In evaluations on real-world deployments, the system demonstrated significant improvements over traditional static placement approaches. By intelligently distributing microservices across available regional resources based on current carbon intensity and pricing, Aceso achieved a 37.4% reduction in carbon emissions and a 3.6% decrease in operational costs while consistently meeting service level objectives (SLOs) for latency-sensitive applications.

The research, published as arXiv:2603.10768, represents a practical approach to sustainable cloud computing for the majority of businesses that lack access to global-scale infrastructure. Aceso's architecture specifically considers the constraints faced by national and regional SMEs, making carbon-aware scheduling accessible to organizations that previously couldn't implement such strategies due to infrastructure limitations or complexity concerns.