Chapter Summary

Chapter Summary

Key Points

• 1.

  Multi-access coded caching generalizes MAN: each user accesses $L$ caches (not just 1). Effective per-user memory $L \cdot M$ if placement is designed so accessed caches are disjoint.

• 2.

  HKD cyclic wrap-around scheme (2017) achieves rate $R_\text{MA}(L) = K(1 - LM/N)/(1 + KLM/N)$ — MAN rate with effective memory $LM$. Gain over single-access: factor $L$.

• 3.

  Cyclic topology matches femtocell / Wi-Fi mesh deployments. In dense urban areas, users naturally see multiple APs — multi-access is the natural model.

• 4.

  Resolvable combinatorial designs extend HKD to non-cyclic topologies. Kirkman triples, affine planes, and similar structures enable multi-access schemes with richer parameter regimes.

• 5.

  CommIT contribution (Wan-Caire 2023) on multi-tier multi-access caching provides a theoretical grounding for 5G MEC deployments. Practical operator collaborations (Orange, TIM, Ericsson) are in research pilot stage.

• 6.

  Tradeoff: coordination vs gain. Multi-access requires coordinated placement across AP cluster. Random (decentralized) placement wastes effective memory. Standards (3GPP Rel-17/18) enable the required coordination.

• 7.

  Stadium example: $L = 4$, $K = 1000$ fans, $\mu = 0.1$: rate reduction $\sim 55\times$ over single-access — enormous practical value for dense-venue Wi-Fi.