Data compression is an efficient technique for saving data storage and transmission costs in networks. Traditional data compression methods usually compress each content item according to its own statistical distribution of symbols and do not take into account user preferences on various content items. However, user preferences significantly impact the statistical distributions of symbols transmitted over communication links. This paper presents an edge source coding method to compress data at the network edge, in which codebooks are designed based on not only the statistical distributions of symbols in the content items but also the user preferences. In edge source coding, multiple content items might be compressed via the same codebook. For discrete user preferences, DCA (difference of convex functions programming algorithm) based and k -means++ based algorithms are proposed to derive codebook designs. For continuous user preferences, a sampling method is applied to yield codebook designs. In addition, edge source coding is extended to the two-user case and codebooks are designed to utilize multicasting opportunities. Simulation results demonstrate that edge source coding significantly reduces transmission costs for short content items.
All Science Journal Classification (ASJC) codes
- Electrical and Electronic Engineering
- Lossless data compression
- codebook design
- difference of convex functions
- edge source coding
- user preference