1. ICLB: intelligent controllers load balancing for software-defined based optical data center networks.
- Author
-
Geresu, Kassahun, Gu, Huaxi, Fadhel, Meaad, Wei, Wenting, and Yu, Xiaoshan
- Subjects
- *
INTELLIGENT control systems , *SERVER farms (Computer network management) , *COMPUTER network traffic , *SOFTWARE-defined networking , *LOAD balancing (Computer networks) , *OPTICAL switches - Abstract
In optical data center networks, the utilization of several software-defined network controllers has been implemented to enhance scalability and reliability. However, during dynamic variations in network traffic, static controller deployment often leads to load imbalance among the controllers. Previous studies have proposed dynamic switch migration as a solution to realize controller load balancing. However, the effectiveness of existing switch migration methods remains insufficient. The load balancing performance of controllers demonstrates limited enhancement following migration, and the migration operations themselves result in supplementary costs. As a result, the issue of switch migration efficiency remains unresolved. To tackle controller load balancing, this research introduces the intelligent controller load balancing (ICLB). ICLB incorporates multiple load metrics to evaluate controller load and selects the ideal target controller based on available resources, leading to enhanced load balancing effectiveness. Moreover, it selects switches while the minimizing migration process, effectively reducing supplementary expenses. Additionally, ICLB can simultaneously manage the workload of multiple controllers during a switch migration process, resulting in enhanced migration efficiency. The outcomes of simulations provide evidence that ICLB offers substantial improvements in controller load balancing performance, along with reduced migration expenses compared to existing solutions. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF