CD-LwTE: Cost-and Delay-aware Light-weight Transcoding at the Edge

The edge computing paradigm brings cloud capabilities close to the clients. Leveraging edge’s can improve video streaming services by employing storage capacity and processing power at for caching transcoding tasks, respectively, resulting in with higher quality lower latency. In this paper, we propose, a Cost-and Delay-aware Light-weight Transcoding approach Edge, context of HTTP Adaptive Streaming (HAS). encoding segment requires computationally intensive search processes. main idea is store optimal results as metadata each bitrate segments reuse it servers reduce required time computational resources transcoding. Aiming minimizing cost delay Video-on-Demand (VoD) services, formulate problem selecting an policy serving requests server, including (i) storing (ii) from (iii) fetching origin or CDN Binary Linear Programming (BLP) model. As result, stores popular serves unpopular ones using origin/CDN server. way, addition significant reduction bandwidth costs, requested remarkably decreased utilizing its corresponding metadata. Moreover, prove proposed BLP model NP-hard propose two heuristic algorithms mitigate complexity . We investigate performance comprehensive scenarios various contents, software, settings, available edge. experimental show that our reduces up 97%, decreases cost, storage, computation, 75%, 48% compared approaches.