Automatic proxy-based watermarking for WWW

With the setting up of digital libraries and information archive centers on Internet, copyright issue is becoming an important issue of concern. Unlike hardcopies, digital data can be modified and distributed very easily. Watermarking is the solution to this copyright problem. By embedding a secure, identifiable mark on the digital data, its data ownership and content integrity can be ensured. Currently, the watermarking process is often done manually in an off-line mode. This makes the updating of watermarking techniques and watermark logo content difficult. Enforcement of watermarking policies for content provisioning on Internet in an organization is also not easy. Furthermore, installing watermarking software on each workstation of an organization is very cost-ineffective. In this paper, we propose a novel solution to address the copyright issue of digital data on Internet. The basic idea is to migrate the watermarking process from content developers into the reverse proxy server that is next to an organization's gateway. Data will be watermarked only when it is retrieved from the organization. The main advantages of this approach are (i) cost effective as it is independent of the amount of data in the web content server and of the number of content developers in an organization, (ii) easy update and maintenance as it is an one-for-all solution, and (iii) effective enforcement as the guideline for watermarking is set in the centralized reverse proxy machine. Both design considerations and implementation details of watermarking in a fully-functional, Squid-based reverse proxy server are presented and its performance is analyzed. All results show that automatic watermarking in a reverse proxy server is a practical, cost and enforcement effective solution to handle copyright issue for digital libraries on Internet.