Optimal Latency-Oriented Coding and Scheduling in Parallel Queuing Systems

The evolution of 5G and Beyond networks has enabled new applications with stringent end-to-end latency requirements, but providing reliable low-latency service high throughput over public wireless is still a significant challenge. One the possible ways to solve this exploit path diversity, encoding information flow multiple streams across parallel links. challenge presented by approach design joint coding scheduling algorithms that adapt state links take full advantage diversity. In paper, we address problem for synchronous traffic source generates data blocks at regular time intervals (e.g., video constant frame rate) needs deliver each block within predetermined deadline. We first develop closed-form performance analysis in simple case two servers without any buffering single-packet blocks, propose model general based on Markov Decision Process (MDP). apply policy iteration obtain maximizes fraction delivered deadline: our simulations show drawbacks different commonly applied heuristic solutions, drawing insights optimal policy.