Collecting statistics from graph-based data is an increasingly studied topic in the data mining community. We argue that they can have great value in the coordination of dynamic IoT systems as well, especially to support complex coordination strategies related to distributed situation recognition.

Thanks to a mapping to the field calculus, a distribution coordination model proposed for collective adaptive systems, we show that many existing “centrality measures” for graphs can be naturally turned into field computations that compute the centrality of nodes in a network. Not only this mapping gives evidence that the field coordination is well-suited to accommodate massively parallel computations over graphs, but also it provides a new basic “brick” of coordination which can be used in several contexts, there including improved leader election or network vulnerabilities detection. We validate our findings by simulation, first measuring the ability of the translated algorithm to self-adjust to network changes, then investigating an application of centrality measures for data summarisation.

从基于图的数据收集统计信息是数据挖掘社区中一个越来越多研究的主题。我们认为，它们在动态物联网系统的协调中也具有重要价值，尤其是对于支持与分布式情况识别相关的复杂协调策略。

多亏了对现场演算的映射，这是针对集体自适应系统提出的分布协调模型，我们表明，许多现有的图“中心测度”可以自然地转化为计算网络中节点中心度的场计算。这种映射不仅证明现场协调非常适合在图形上进行大规模并行计算，而且还提供了一种新的基本“协调”方法，可以在多种情况下使用，其中包括改进的领导者选举或网络漏洞检测。我们通过仿真验证了我们的发现，首先测量了转换后的算法针对网络变化进行自我调整的能力，然后研究了数据汇总的集中度测量的应用。