Cost-Effective Resource Configuration for Cloud Video Streaming Services

Video streaming services are migrating to cloud environments for the economic expense with good scalability. However, cloud providers offer flexible resource configurations, e.g., on-demand, reserved and spot instances, with significant different pricing policies, of which one single configuration is suboptimal for cloud video streaming services. In this paper, we propose hybrid configuration schemes of cloud video streaming services to reduce the cost. To achieve this goal, we first introduce a lightweight prediction algorithm to predict the future video traffic. With the predicted video traffic, we then give the Hybrid-R hybrid configuration scheme by configuring both on-demand and reserved instances, and the Hybrid-RS hybrid configuration scheme by further configuring spot instances. Our evaluations using traces from real video service providers show that our configuration schemes can reduce cost by at least 20% compared to the unoptimized ones with negligible overhead.