Modeling the Value of JITEX: Just-In-Time Transcoding at the Edge

Abstract

Content delivery networks (CDNs) have developed virtualized technologies and 5G data centers to enable cost-effective edge computing. Edge processing overcomes cache limitations by executing compute-intensive tasks at or near edge servers. It is expected to fix some issues related to adaptive bitrate (ABR) over-the-top (OTT) streaming. In particular, just-in-time-transcoding at the edge (JITEX) allows the edge server to cache a single pivot video and to offer nonetheless multiple video versions. This article models the impact of JITEX by examining various cost variables, use cases, and CDN implementation. For the sake of transparency, we use public data and well-known popularity models. By sharing this JITEX model, we aim to facilitate discussion on deploying transcoders at the edge, dispelling misconceptions, and identifying profitable research areas.