Image Abstraction Using Anisotropic Diffusion Symmetric Nearest Neighbor Filter

Abstract

Currently, more and more Internet of Things (IoT) applications use Internet cameras to sense the physical world, and a large amount of streaming media data is uploaded onto video data center. Consequently, it brings much pressure on the media streams' caching and storage resources of data centers. For solving this problem, we propose a streaming media caching resource scheduling approach which gives full consideration on the resource constraints of cameras and the caching server instability. We adopt a server-slave model to manage the caching server cluster. The management server monitors the state of each slave caching server in real-time, and schedules the connecting requests of Internet cameras according to the stability of the distributed caching servers. Therefore, we can obtain the higher reliability for streaming media uploading and the better workload balance among the caching servers. Based on the proposed scheduling approach, we implement a distributed caching system for reliable video uploading in data center, and the experimental results validate the effectiveness of our approach. © 2012 Springer-Verlag.