Today, millions of physical devices are visible on the Internet through IP addresses, such as wearables, residential routers, cameras and industrial control devices. Users can access and control online devices, while attackers also target them, bringing potential security concerns. In this article, we propose a conceptual framework for discovering, recognizing and managing these online devices at the Internet-wide scale. It is the first work of Internet-wide device search, including three modules: network measurement, fingerprinting, and applications. We have implemented a prototype system and real-world experiments to illustrate the effectiveness of the framework. Furthermore, we explore the use and application of the system through searching camera devices in the whole IPv4 space. We deployed the system on Amazon EC2 to search for webcams and ICS devices nine times from April 2015 to April 2016, and we found 1.3 million webcams in total. In a comprehensive analysis of our dataset, we have characterized the use of online physical devices on the Internet.

中文翻译：

---

搜索互联网设备的框架

如今，数以百万计的物理设备通过IP地址在Internet上可见，例如可穿戴设备，住宅路由器，摄像机和工业控制设备。用户可以访问和控制在线设备，而攻击者还可以将其作为目标，从而带来潜在的安全隐患。在本文中，我们提出了一个概念框架，用于在Internet范围内发现，识别和管理这些在线设备。这是Internet范围内设备搜索的第一项工作，包括三个模块：网络测量，指纹识别和应用程序。我们已经实现了原型系统和真实世界的实验，以说明该框架的有效性。此外，我们通过在整个IPv4空间中搜索摄像头设备来探索该系统的使用和应用。从2015年4月到2016年4月，我们在Amazon EC2上部署了该系统以搜索网络摄像头和ICS设备九次，总共发现了130万个网络摄像头。在对我们的数据集进行全面分析后，我们对互联网上在线物理设备的使用进行了特征化描述。