Critter@home: Content-Rich Traffic Trace Repository from Real-Time, Anonymous, User Contributions
ABSTRACT
In this project we investigate, build and deploy a publicly accessible repository of up-to-date application-level data. This data is sorely needed in the networking and security communities. Currently available network data with application-level information is often outdated and is either private or customized to specific, narrow research needs. We will address this problem by designing and deploying a publicly accessible repository of application-level data called Critter- at-Home, where Critter stands for Content-Rich Traffic Trace Repository. We have the following design goals for Critter-at-Home:
-
Data is fresh and diverse. We will achieve this goal by creating an anonymous contributor network where contributors can join and leave at will, ensuring continuous data collection. Our framework will be flexible and allow for a broad range of contribution modes, thus attracting diverse participants.
-
Data utility is maximized. We will achieve this by minimally processing contributed data and allowing researchers to query this "almost raw" data for features that interest them.
-
Contributor privacy and identity are fully protected. We will achieve this by giving contributors full control over their data and its usage at a fine-grain level including mechanisms to withdraw data fully, store data remotely or locally, and contribute only what they are comfortable with. All stored data will first be sanitized to alter personal and private information and then encrypted. Our secure query framework will protect privacy further by allowing only aggregate and prevalent results to be returned to researchers, thus protecting data contributor privacy by using the "hiding in a crowd" approach.
This research addresses the security and privacy focus of the Trustworthy Computing solicitation in two ways. First, our work will enable access to content-rich network data, which is essential to continued progress in networking and cybersecurity research. Second, our work will explore new approaches to secure sharing of private data, which is a recurring problem in many facets of networking and cybersecurity.
Project ID: 1224035
- PDF document
- 4.2 MB
- 121 downloads
- Download
- PDF version
- Printer-friendly version