Hive oversight for network intrusion early warning using DIAMoND: a bee-inspired method for fully distributed cyber defense
Hive oversight for network intrusion early warning using DIAMoND: a bee-inspired method for fully distributed cyber defense
复制标题
使用 DIAMoND 进行网络入侵预警的 Hive 监督：一种受蜜蜂启发的完全分布式网络防御方法
DOI：
--
复制
发表时间：
2016
影响因子：
11.2
通讯作者：
N. Fefferman
中科院分区：
文献类型：
--
作者：
Maciej Korczyński;Ali Hamieh;Jun Ho Huh;Henrik Holm;S. Raj Rajagopalan;N. Fefferman
社会性昆虫群体在进化过程中得以生存，部分原因是它们的合作方法取得了成功：使用本地信息和分布式决策算法来检测和利用环境中的关键资源。这些方法具有不寻常的和有用的能力，可以快速检测异常，只需很少的内存，并且只使用非常本地的信息。我们的研究探讨了自组织异常检测系统的潜力，该系统的灵感来自于蜜蜂群体中自然观察到的异常。我们提供了一个总结的结果，从最近提出的算法的非参数，完全分布式的协调框架，这些方法的生物成功转化为类似的操作用于网络防御和讨论的功能，启发了这种翻译。我们探讨的影响，检测性能的定义范围内的分布式通信的每个节点，只涉及一小部分的总节点在网络中的分布式检测通信。我们评估我们的算法使用基于软件的测试实现，并证明了高达20%的改进，在并行隔离异常检测器的检测能力。
Social insect colonies have survived over evolutionary time in part due to the success of their collaborative methods: using local information and distributed decision making algorithms to detect and exploit critical resources in their environment. These methods have the unusual and useful ability to detect anomalies rapidly, with very little memory, and using only very local information. Our research investigates the potential for a self-organizing anomaly detection system inspired by those observed naturally in colonies of honey bees. We provide a summary of findings from a recently presented algorithm for a nonparametric, fully distributed coordination framework that translates the biological success of these methods into analogous operations for use in cyber defense and discuss the features that inspired this translation. We explore the impacts on detection performance of the defined range of distributed communication for each node and of involving only a small percentage of total nodes in the network in the distributed detection communication. We evaluate our algorithm using a software-based testing implementation, and demonstrate up to 20 percent improvement in detection capability over parallel isolated anomaly detectors.