A carregar...
Publicação
Paprika
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.11 MB | Adobe PDF |
Orientador(es)
Resumo(s)
Efficient data replica placement at the network edge is crucial for minimizing data access latency and enhancing user experience. Current solutions are mostly popularity oriented, assuming that users in close proximity share similar interests. While this strategy is effective for widely accessed data, it overlooks the specificity of user interests and their affinities, especially in diverse environments managed by different edge servers: data popular in region A may not be relevant in region B but be highly valuable in region C.We introduce Paprika, an online heuristic-based hybrid algorithm that combines the strengths of Genetic Algorithms and Tabu Search to address replica selection and placement in edge computing environments. Paprika takes into account data popularity, pair-wise affinity between regions, and server storage capacity. Our evaluation demonstrates that the hybrid approach outperforms traditional heuristic methods by better balancing user interests across regions and favoring regions with stronger affinities.
Descrição
Funding Information:
This work is supported by NOVA LINCS (UIDB/04516/2020) with the financial support of FCT.IP.
Publisher Copyright:
Copyright © 2025 held by the owner/author(s).
Palavras-chave
Affinity Capacity Edge computing Popularity Replica placement Software
Contexto Educativo
Citação
Editora
ACM - Association for Computing Machinery