多数のセンサーが送出するトラヒックに対するアプリケーション分析手法 (インターネットアーキテクチャ)

We discuss an anomaly detection of a computer network in Big data era. The network has to support to deliver network traffic which is sent from many devices such as a large number of sensors and smartphones. The core network in Big data era is required to have a feature of anomaly detection for many devices. We propose a way to create traffic patterns from such kind of devices by using NMF (non-negative matrix factorization). The analysis enable to make clusters with the same traffic behavior and to show their traffic patterns immediately.