| _version_ | 1866901945664405504 |
|---|---|
| author | Aslan, Hacı İsmail Witzke, Joel Murphy, Amy Lynn Wang, Yiming Berasi, Davide Markus, Andras |
| author_facet | Aslan, Hacı İsmail Witzke, Joel Murphy, Amy Lynn Wang, Yiming Berasi, Davide Markus, Andras |
| contents | <p>This deliverable presents our work on resource selection based on the “logical proximity” notion and extends it by proposing strategies to orchestrate microservices in Cloud-to-Edge continuum, designed to support scalable, self-organizing deployment of microservices across the Cloud-to-Edge continuum. This deliverable presents a decentralized orchestration model developed under the Swarmchestrate project, inspired by swarm intelligence and designed to support scalable, self-organizing deployment of microservices across the Cloud-to-Edge continuum. At its core is the novel use of the logical proximity concept, which enables distributed matchmaking between application requirements and infrastructure capabilities. By quantifying resource suitability using multi-criteria cost functions and ranking methods such as Borda voting, the system autonomously forms optimal deployment groups (referred to as "swarms") that satisfy QoS goals like low latency, minimal energy use, and reduced<br>cost. </p> <p>To evaluate the feasibility and effectiveness of this approach, a full-stack prototype was developed and tested. Key achievements include:</p> <ul> <li>Design and implementation of a decentralized matchmaking algorithm using logical proximity and multi-objective QoS ranking.</li> <li>Development of a clean, modular orchestration framework based on containerized microservices and Kubernetes/K3s.</li> <li>Simulation-based validation of the matchmaking strategy using DISSECT-CF-Fog, showing effective swarm formation and resource selection.</li> <li>Robust fault-tolerant control plane based on Raft consensus for leadership among RLAs.</li> <li>Design of a system for runtime-optimization using AI techniques.</li> </ul> <p>These contributions lay the groundwork for a self-adaptive, resilient orchestration layer that removes the need for centralized control or manual intervention during deployment. The system is designed to evolve toward runtime QoS adaptation and integration with broader orchestration mechanisms, making it a foundational element of the overall Swarmchestrate platform.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_17101558 |
| institution | Zenodo |
| language | eng |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | D1.1 - Logical Proximity and Distributed Matchmaking Algorithms Aslan, Hacı İsmail Witzke, Joel Murphy, Amy Lynn Wang, Yiming Berasi, Davide Markus, Andras <p>This deliverable presents our work on resource selection based on the “logical proximity” notion and extends it by proposing strategies to orchestrate microservices in Cloud-to-Edge continuum, designed to support scalable, self-organizing deployment of microservices across the Cloud-to-Edge continuum. This deliverable presents a decentralized orchestration model developed under the Swarmchestrate project, inspired by swarm intelligence and designed to support scalable, self-organizing deployment of microservices across the Cloud-to-Edge continuum. At its core is the novel use of the logical proximity concept, which enables distributed matchmaking between application requirements and infrastructure capabilities. By quantifying resource suitability using multi-criteria cost functions and ranking methods such as Borda voting, the system autonomously forms optimal deployment groups (referred to as "swarms") that satisfy QoS goals like low latency, minimal energy use, and reduced<br>cost. </p> <p>To evaluate the feasibility and effectiveness of this approach, a full-stack prototype was developed and tested. Key achievements include:</p> <ul> <li>Design and implementation of a decentralized matchmaking algorithm using logical proximity and multi-objective QoS ranking.</li> <li>Development of a clean, modular orchestration framework based on containerized microservices and Kubernetes/K3s.</li> <li>Simulation-based validation of the matchmaking strategy using DISSECT-CF-Fog, showing effective swarm formation and resource selection.</li> <li>Robust fault-tolerant control plane based on Raft consensus for leadership among RLAs.</li> <li>Design of a system for runtime-optimization using AI techniques.</li> </ul> <p>These contributions lay the groundwork for a self-adaptive, resilient orchestration layer that removes the need for centralized control or manual intervention during deployment. The system is designed to evolve toward runtime QoS adaptation and integration with broader orchestration mechanisms, making it a foundational element of the overall Swarmchestrate platform.</p> |
| title | D1.1 - Logical Proximity and Distributed Matchmaking Algorithms |
| url | https://doi.org/10.5281/zenodo.17101558 |