Szczegóły publikacji
Opis bibliograficzny
Enhanced Lagrange Decomposition for multi-objective scalable TE in SDN / Piotr JAGLARZ, Piotr BORYŁO, Andrzej SZYMAŃSKI, Piotr CHOŁDA // Computer Networks ; ISSN 1389-1286. — 2020 — vol. 167 art. no. 106992, s. 1–11. — Bibliogr. s. 10–11, Abstr. — Publikacja dostępna online od: 2019-11-16
Autorzy (4)
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 126448 |
|---|---|
| Data dodania do BaDAP | 2020-01-14 |
| Tekst źródłowy | URL |
| DOI | 10.1016/j.comnet.2019.106992 |
| Rok publikacji | 2020 |
| Typ publikacji | artykuł w czasopiśmie |
| Otwarty dostęp |  |
| Czasopismo/seria | Computer Networks |
Abstract
The paradigm of Software Defined Networking opens attractive perspectives for network operators in terms of traffic engineering (TE) mechanisms. Thanks to programmability of a logically centralized network controller, static optimization techniques became applicable to optimization of resource utilization. Such an approach, combined with accurate traffic prediction, enabled deployment of globally scoped, efficient solutions. However, such an approach might pose some scalability issues which must be carefully considered. The challenge is to propose a solution able to solve multi-objective problems with non-linear constraints in large-scale networks. Networks, that are comprised of numerous nodes handling millions of flows of dynamic nature. Our work is aimed at addressing these issues. A novel, energy-aware, multi-objective, mixed integer linear programming problem is formulated, and solved using the Lagrange Decomposition method. The method is enhanced by novel adoption of ergodic sequences in order to improve quality of a primal solution recovered from a dual one. The proposed heuristics is carefully assessed regarding its time-efficiency, applicability to multi-objective optimization problems and problems with constraints of non-linear nature, such as energy consumption. The presented method is of significant practical value, and as such is a direct response to traffic engineering scalability challenges in Software Defined Networks. © 2019 Elsevier B.V.
