可编程网络数据平面技术进展

网络数据平面执行数据包处理转发,是网络性能的决定性因素之一。大带宽、低时延、可持续演进的网络基础设施需要构建高效可编程的网络数据平面。首先,介绍数据包处理转发模型,并以此为基础概述网络数据平面在性能与可编程性面临的关键挑战。然后,从数据包查找算法理论与软/硬件协同实现机制出发,详细论述其基本思路及关键核心技术进展以应对上述关键挑战。最后,探讨高效可编程数据平面的未来发展趋势与技术演进路线。

Dynamic spectrum allocations in multi-cells and intra-cell of cognitive network to enhance system performance

In this study, we propose new dynamic spectrum allocations in multi-cells and intra-cell of cognitive network to enhance system performance in terms of decreasing probability of interruption and spectrum handoff of communication services in a cognitive system. The inter-cells of the spectrum allocation mechanism is designed to share the risk of vacating spectrum caused by licensed incumbents re-occupying the spectrum and minimize probability of service interruption in the cognitive network. This mechanism also can guarantee fairness among multi-cells. The intra-cell of the proposed spectrum allocation is based on a service data hierarchical model and establishes a mapping mechanism between layered data and the spectrum. It can reduce probability of spectrum handoff. Finally, simulation results are given and show that the new mechanism can reduce service interruption ratio and the probability of spectrum handoff caused by licensed incumbents with re-occupying the spectrum.