# Cache-Aided Interference Management in Partially Connected Linear Networks

IEEE Transactions on Communications, pp. 301-316, 2020.

EI WOS

Weibo:

Abstract:

This paper studies caching in
<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$(K\!+\!L\!-\!1)\!\times \!K$ </tex-math></inline-formula>
partially connected wireless linear networks, where each of the
<inline-formula xmlns:mml="http://www.w3.org/1998/...More

Code:

Data:

Introduction

- Caching is a novel solution to enhance the communication efficiency by exploiting the increasingly rich storage resource in wireless networks.
- This is because the connected receiver set of transmitter j in the circular network, denoted as Rcj = {j − L + 1, j − L + 2, .
- J} mod K, is the union of the connected receiver sets of transmitters j and K +j in the linear network, i.e., Rcj = Rj ∪ RK+j.

Highlights

- Caching is a novel solution to enhance the communication efficiency by exploiting the increasingly rich storage resource in wireless networks
- The authors in [4] showed that the original interference channel can be turned into more favorable channels including X channel and broadcast channel by proper file splitting and placement
- We present a new file splitting and cache placement strategy, which is modified from the novel scheme in [8] by taking the partial network connectivity into account
- Similar to Example 1, to ensure that each transmitter is connected to L receivers, we introduce virtual receivers to transform the network into the expanded partially connected linear interference networks
- We present a lower bound of the minimum normalized delivery time, followed by the discussion on the optimality of the achievable scheme presented in the previous sections
- The multiplicative gap between the achievable upper bound and the theoretical lower bound of the minimum normalized delivery time for the considered network is less than 2

Results

- Each transmitter j has an encoding function Λj to map its cached content Uj, receiver demand d, and network-wide channel realization H = [hij (t)]i∈[K−1]+,j∈Ri,t∈[T ] to the signal (Xj[t])Tt=1 Λj(Uj, d, H), where T is the block length of the code.
- The subfile Wn,1,{0,1}, which is cached in transmitters {1, 4} and receivers {0, 1, 3} in the 6 × 4 network in Fig. 1, has a size of a2F bits.
- Note that transmitter 0 and 5 only have two coded messages to send, because the virtual receiver groups {−2, −1} and {4, 5} do not request any files.
- The network topology is converted to an expanded partially connected Xmulticast channel with multicast size 2, where every set of two receivers form a multicast group, and desire a common message from any of the transmitters that are connected to both of them.
- Denote vj{i1,i2} as the transmit beamforming vector of message Wt{ji1,i2} and Hij as the channel realization between transmitter j and receiver i.
- Based on Algorithm 1, the network topology is converted into the expanded partially connected X-multicast channel with multicast size r + 1, where each transmitter j has an independent coded message WtRj intended to the actual receivers in multicast group R satisfying
- (Achievable NDT) For the cache-aided (K + L − 1) × K partially connected linear interference network, an achievable NDT is given by the optimal solution of the following linear programming (LP) problem: τ ∗ ≤τub ar s.t.

Conclusion

- The expression in (7) suggests that the proposed scheme III, the cached contents in the transmitter pair (j, K + j) of achieves the receiver local caching gain of (1 − μR) and the linear network will be identical when L is a divisor of a combined coded multicasting and transmitter coordination gain of
- L − 1) × K partially connected linear interference network, the minimum NDT is lower-bounded by: APPENDIX A: PROOF OF LEMMA 1 τ ∗ ≥ 1 − μR.
- Each transmitter j has an independent message, denoted as WtRj , aimed for the actual receivers in each of its connected receiver multicast group R satisfying

Summary

- Caching is a novel solution to enhance the communication efficiency by exploiting the increasingly rich storage resource in wireless networks.
- This is because the connected receiver set of transmitter j in the circular network, denoted as Rcj = {j − L + 1, j − L + 2, .
- J} mod K, is the union of the connected receiver sets of transmitters j and K +j in the linear network, i.e., Rcj = Rj ∪ RK+j.
- Each transmitter j has an encoding function Λj to map its cached content Uj, receiver demand d, and network-wide channel realization H = [hij (t)]i∈[K−1]+,j∈Ri,t∈[T ] to the signal (Xj[t])Tt=1 Λj(Uj, d, H), where T is the block length of the code.
- The subfile Wn,1,{0,1}, which is cached in transmitters {1, 4} and receivers {0, 1, 3} in the 6 × 4 network in Fig. 1, has a size of a2F bits.
- Note that transmitter 0 and 5 only have two coded messages to send, because the virtual receiver groups {−2, −1} and {4, 5} do not request any files.
- The network topology is converted to an expanded partially connected Xmulticast channel with multicast size 2, where every set of two receivers form a multicast group, and desire a common message from any of the transmitters that are connected to both of them.
- Denote vj{i1,i2} as the transmit beamforming vector of message Wt{ji1,i2} and Hij as the channel realization between transmitter j and receiver i.
- Based on Algorithm 1, the network topology is converted into the expanded partially connected X-multicast channel with multicast size r + 1, where each transmitter j has an independent coded message WtRj intended to the actual receivers in multicast group R satisfying
- (Achievable NDT) For the cache-aided (K + L − 1) × K partially connected linear interference network, an achievable NDT is given by the optimal solution of the following linear programming (LP) problem: τ ∗ ≤τub ar s.t.
- The expression in (7) suggests that the proposed scheme III, the cached contents in the transmitter pair (j, K + j) of achieves the receiver local caching gain of (1 − μR) and the linear network will be identical when L is a divisor of a combined coded multicasting and transmitter coordination gain of
- L − 1) × K partially connected linear interference network, the minimum NDT is lower-bounded by: APPENDIX A: PROOF OF LEMMA 1 τ ∗ ≥ 1 − μR.
- Each transmitter j has an independent message, denoted as WtRj , aimed for the actual receivers in each of its connected receiver multicast group R satisfying

- Table1: Subfiles to be delivered with r = |Q| = 1

Funding

- This work is supported by the National Natural Science Foundation of China under grants 61571299, 61329101, and 61521062

Reference

- H. Liu, Z. Chen, X. Tian, X. Wang, and M. Tao, “On content-centric wireless delivery networks,” IEEE Wireless Commun., vol. 21, no. 6, pp. 118–125, December 2014.
- N. Golrezaei, A. F. Molisch, A. G. Dimakis, and G. Caire, “Femtocaching and device-to-device collaboration: A new architecture for wireless video distribution,” IEEE Commun. Magazine, vol. 51, no. 4, pp. 142–149, April 2013.
- M. A. Maddah-Ali and U. Niesen, “Fundamental limits of caching,” IEEE Trans. Information Theory, vol. 60, no. 5, pp. 2856–2867, May 2014.
- M. Maddah-Ali and U. Niesen, “Cache-aided interference channels,” in IEEE ISIT, June 2015.
- A. Sengupta, R. Tandon, and O. Simeone, “Cloud and cache-aided wireless networks: Fundamental latency trade-offs,” 2016. [Online]. Available: http://arxiv.org/abs/1605.01690
- X. Yi and G. Caire, “Topological coded caching,” in IEEE ISIT, July 2016.
- F. Xu, K. Liu, and M. Tao, “Cooperative Tx/Rx caching in interference channels: A storage-latency tradeoff study,” in IEEE ISIT, July 2016.
- F. Xu, M. Tao, and K. Liu, “Fundamental tradeoff between storage and latency in cache-aided wireless interference networks,” to appear in IEEE Trans. Information Theory, DOI: 10.1109/TIT.2017.2717912.
- J. Hachem, U. Niesen, and S. N. Diggavi, “Degrees of freedom of cache-aided wireless interference networks,” 2016. [Online]. Available: http://arxiv.org/abs/1606.03175
- N. Naderializadeh, M. A. Maddah-Ali, and A. S. Avestimehr, “Fundamental limits of cache-aided interference management,” IEEE Trans. Information Theory, vol. 63, no. 5, pp. 3092–3107, May 2017.
- Y. Cao, M. Tao, F. Xu, and K. Liu, “Fundamental storage-latency tradeoff in cache-aided MIMO interference networks,” IEEE Trans. Wireless Commun., vol. 16, no. 8, pp. 5061–5076, August 2017.
- M. Wigger, R. Timo, and S. Shamai, “Complete interference mitigation through receiver-caching in Wyner’s networks,” in IEEE ITW, September 2016.
- A. E. Gamal, V. S. Annapureddy, and V. V. Veeravalli, “Interference channels with coordinated multipoint transmission: Degrees of freedom, message assignment, and fractional reuse,” IEEE Trans. Information Theory, vol. 60, no. 6, pp. 3483–3498, June 2014.
- V. Cadambe and S. Jafar, “Interference alignment and the degrees of freedom of wireless X networks,” IEEE Trans. Information Theory, vol. 55, no. 9, pp. 3893–3908, September 2009.
- M. Zamanighomi and Z. Wang, “Degrees of freedom region of wireless x networks based on real interference alignment,” IEEE Trans. Information Theory, vol. 62, no. 4, pp. 1931–1941, April 2016.

Full Text

Tags

Comments