Back-to-Back Butterfly Network, an Adaptive Permutation Network for New Communication Standards

Hassan Harb (1, 2) , Cyrille Chavet (3, 4, 1)

1 Lab-STICC_UBS_CACS_MOCS
2 Lab-STICC - Laboratoire des sciences et techniques de l'information, de la communication et de la connaissance
3 Lab-STICC_UBS_CACS_MOCS
4 Lab-STICC - Laboratoire des sciences et techniques de l'information, de la communication et de la connaissance

In this paper, we introduce an adaptive Back-to-Back Butterfly Network (B 2 BN) dedicated to next communication standards. It can perform any kind of permutation, and its architecture is based on a concatenation of basic networks. However for a set of permutations, the selection of non-conflicting paths is still a complex task. Hence, in our paper we rely on the properties of the proposed architecture, to introduce a formal model that efficiently solve such conflicts. From the collection of all possible paths in the targeted B 2 BN , the proposed method select the conflict-free paths to transfer data from the input to their permuted output (w.r.t., a ad hoc constraints model). Once the appropriate paths are selected, the control signals for B 2 BN are generated. This model has been experimented with 5G communication, showing how to process several frames in parallel with different permutation constraints.

Mots clés: Index Terms-Butterfly Network; Back-to-Back Butterfly Network; Permutation; QC-LDPC; 5G

Domaines: Architectures Matérielles [cs.AR]; Systèmes embarqués; Traitement du signal et de l'image [eess.SP]; Ingénierie assistée par ordinateur; Algorithme et structure de données [cs.DS]

Hassan Harb, Cyrille Chavet. Back-to-Back Butterfly Network, an Adaptive Permutation Network for New Communication Standards. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), May 2020, Barcelonne, Spain. ⟨hal-02493104⟩