EFM: Improving DCNs throughput using the transmission rates of elephant flows

Rocha, Andre; Verdi, Fábio L.

doi:10.1109/iscc.2018.8538633

Cited by 4 publications

(1 citation statement)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Elephant traffic accounts for 10% of the total traffic, but carries 90% of the data volume. Because ECMP cannot dynamically obtain network data, multiple elephant flows may be allocated to the same path, resulting in link congestion and affecting network performance, so this method is not suitable for current data center networks [4]. The ACO algorithm is a heuristic global optimization algorithm.…”

Section: Introductionmentioning

confidence: 99%

Research on data center load balancing based on particle swarm optimization fusion ant colony optimization algorithm

Song¹,

Jun²

2023

3rd International Conference on Applied Mathematics, Modelling, and Intelligent Computing (CAMMIC 2023)

View full text Add to dashboard Cite

In recent years, with the continuous development of big data and cloud computing services, the scale of data centers has become larger and larger, and the rapidly increasing network traffic has also put forward higher requirements for the load balancing technology of data centers. Most of the traditional traffic scheduling algorithms are derived from IP networks. These algorithms do not take into account performance parameters such as link bandwidth and delay of real networks, resulting in very unsatisfactory results. In view of the above problems, this paper proposes a traffic scheduling algorithm based on the Particle Swarm Optimization Fusion Ant Colony Optimization (P-ACO) algorithm, which is mainly for dynamic traffic scheduling of elephant flow. First, the Software Defined Networking (SDN) controller obtains the network performance parameters and topology, then uses the PSO algorithm to iteratively search for solutions, integrates the solutions into the initial pheromone distribution of the ACO algorithm, and finally combined with the global network state, the optimal solution is obtained through the improved ACO algorithm.

show abstract

Section: Introductionmentioning

confidence: 99%