Stream slip

流媒体碎片

The cleverest part, though, is that, unlike other efforts to exploit leaky video streams, it does not actually need direct access to the stream itself, or even to the device the video is being shown on. By planting a small amount of JavaScript code in a web browser on a personal computer or smartphone that is merely attached to the same Wi-Fi network as the viewer’s device, the film being watched can be identified with almost the same accuracy.

最精妙的部分在于，区别于其他方法需要分析泄露的流媒体内容，这种方式并不需要直接访问流媒体内容本身，甚至是不需要访问播放流媒体的设备。通过在个人电脑或者智能手机的浏览器中植入一小段 JavaScript 代码，就可以知道使用同一 Wi-Fi 网络的用户的设备上所观看的内容，并且可以达到差不多高的识别率。

Web browsers confine JavaScript—which is ubiquitous in web pages and advertisements, and runs automatically—to a“sandbox” supposed to prevent it from collecting private information. JavaScript code can, however, still communicate with the computer server that sent it—and this is enough for Dr Tromer. It enables his implant to flood the entire Wi-Fi network with random data, creating congestion. The result is that a video stream feeding another device on the network will create bursty delays in the JavaScript’s communications with its own server. Measuring these is enough for the spyware to be able to identify the film being watched.

JavaScript 几乎出现于所有的网络页面以及广告之中，网络浏览器通过将 JavaScript 代码放入“沙箱”中来保护用户的个人隐私数据不被泄露。但是 JavaScript 代码仍然会和发送它服务器进行通信，对 Tromer 而言这就足够了。这能帮助他在植入代码以使得整个 Wi-Fi 网络中充满随机数据来制造拥堵。结果就是播放流媒体的设备会在 JavaScript 和服务器的通信过程中知道大量的延迟。通过间谍软件来观测这些数据就能够确认所观看的电影了。