• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International Conference on Advances in Big Data, Computing and Data Communication Systems >Feature Selection Strategies for the Controller Placement Problem in SDNs: A Review
【24h】

Feature Selection Strategies for the Controller Placement Problem in SDNs: A Review

机译：SDNS中控制器放置问题的功能选择策略：审查

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Recent advances in computing makes it possible for a myriad of devices, people and things to connect together irrespective of space and time constraints. However, for this to happen next-generation networks need to be fast, scalable, reliable and cost effective. Interestingly, Software defined networking (SDN) provides the enabling paradigm for this to happen. However, the placement of controllers in this emerging paradigm is crucial to its performance and scalability hence its relevance. To this end, we first present a concise review of four controller placement models relevant to this study, highlighting their objectives, strengths and gaps. We further illustrate how the feature selection strategies used by these models affect their accuracy and convergence times, thus making them unsuitable for real life scenarios. Although three out of four of the reviewed models consider traffic conditions in placing controllers in SDNs, they nonetheless incur high convergence times. There is, however the possibility of achieving high accuracy with low convergence by fine tuning the feature selection techniques implemented by these models. The deliverable in view is a road map towards the development of an optimized controller placement model where optimality means maximizing the model's accuracy whilst minimizing its convergence time. In this way we will be providing the necessary mechanism for an optimized controller placement model that responds to changing traffic conditions.
机译：最近的计算进步使得无数的设备，人物和事物可以连接在一起，而不管空间和时间限制如何。但是，为了实现下一代网络，需要快速，可扩展，可靠和成本效益。有趣的是，软件定义的网络（SDN）为此提供了实现的启用范例。然而，控制器在这种新兴范式中的放置对于其性能和可扩展性至关重要，因此其相关性。为此，我们首先提出了一个简明的审查与本研究相关的四个控制器放置模型，突出了其目标，优势和差距。我们进一步说明了这些模型使用的特征选择策略如何影响其准确性和收敛时间，从而使它们不适合现实生活场景。虽然四种综述模型中的三个考虑在SDNS中放置控制器时的流量条件，但它们仍然会产生高收敛时间。然而，通过精细调整这些模型实现的特征选择技术，通过微调来实现高精度的高精度。可交付的视图是朝向开发优化的控制器放置模型的路线图，其中最佳性意味着最大化模型的准确性，同时最大限度地减少其收敛时间。通过这种方式，我们将为优化的控制器放置模型提供必要的机制，该模型响应不断变化的交通状况。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号