A fast data synchronization technique in optical inter-connected data centers

Lü, Yang; Cao, Lufang; Xu, Kun; Li, Xuemeng; Deng, Hongsheng; Lv, Yunxin; Zhai, Yanrong

doi:10.1016/j.osn.2022.100698

Cited by 1 publication

(1 citation statement)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to the massive volume of data flows, the development of innovative data center architectures and networking technology, which can provide high throughput and low delay with a small cost, is in high demand [4]- [6]. As a result, with the emerging optical technologies, the reconfigurable circuit switch (RCS), which can establish circuit connections among Top-of-Rack (ToR) switches, has been proposed as a promising paradigm for datacenter networks [7]- [10]. Compared to the traditional electronic packet switching pattern, the data can be directly forwarded to the destination rack through setting up new circuits in RCS-based networks.…”

Section: Introductionmentioning

confidence: 99%

Towards Real-Time Non-Preemptive Multicast Scheduling in Reconfigurable Data Center Networks

Zhang,

Liu,

et al. 2024

Preprint

View full text Add to dashboard Cite

Multicast has emerged as a primary communication pattern in datacenter networks due to the increasing demand for distributed data-parallel applications. To accelerate multicast traffic, the emerging reconfigurable circuit technology, which can establish circuit connections among switches, has been proposed as a promising paradigm for datacenter networks. This paper investigates how to accelerate the non-preemptive multicast flows in a demand-aware manner in reconfigurable datacenter networks. Firstly, we introduce the problem of scheduling circuit switches to minimize the average completion time and prove it to be NP-hard. To address the conflicts between different multicast flows under the bandwidth constraint, we consider the concept of splitting multicast flows and propose a two round matching algorithm. Additionally, to further reduce the average completion time, we introduce a method to utilize the remaining capacity of the ToR switches by splitting the unscheduled flows. The proposed algorithm is proved to have an approximation ratio of $2\sqrt{2n}$, where $n$ represents the number of Top-of-Rack (ToR) switches. Finally, the extensive simulations demonstrate the effectiveness of the proposed algorithm in reducing the average completion time of flows compared to state-of-the-art algorithms.

show abstract