Abstract :
The increasing transmission of illegal videos
over the Internet imposes the needs to develop large-scale
digital video forensics systems for prosecuting and
deterring digital crimes in the Internet. In this paper, we
propose, design, and implement a novel large-scale Digital
Forensics Service Platform (DFSP) that can effectively
detect illegal content from Internet videos. More
specifically, we propose a distributed architecture by taking
advantage of Content Delivery Network (CDN) to improve
scalability, which can process enormous number of Internet
videos in real time. We propose CDN-based ResourceAware Scheduling (CRAS) algorithm, which schedules the
tasks efficiently in the DFSP according to resource
parameters, such as delay and computation load. We
deploy the DFSP system in the Internet, which integrates
the CDN-based distributed architecture and CRAS
algorithm with a large-scale video detection algorithm, and
evaluate the deployed system. Our evaluation results
demonstrate the effectiveness of the platform.
Keyword :
Content delivery network, digital forensics, load balancing, Resource scheduling, video detection