Thread by @fgrosshans

Now at #IWQNS, Patrick Thiran from @EPFL presents an overview of classical an and complex networks. Mainly on classical networks, and on tools who might be applied to quantum networks.
#LTQI

@EPFL Patrick Thiran: Networks are complex. Layer abstraction helps to “divide and conquer”, but this abstraction comes at a cost, and sometimes it not worth it.
#LTQI #IWQNS

@EPFL Patrick Thiron: Case study1: wireless multi hop networks.
N→∞ nodes scatterred. We want full connectivity, study throughput scaling.
Layer 1. : Physical layer: adjust network for full connectivity
Layer 2: MAC connectivity
#LTQI #IWQNS

@EPFL Patrick Thiran : Increasing power, long range connectivity appears much before full connectivity, because of percolation. (Nodes at random on the plane)
Percolation (thermodynamic thr): r_n∝1/√n
Full connectivity threshold r_n∝√(log n /n)
Energy ∝r²
#LTQI #IWQNS

@EPFL Patrick Thiran: The MAC layer to schedule and routing needs to have some extra complexity to take into account the few disconnected noise.
#IWQNS #LTQI

*Nodes, not noise
#Erratum #LTQI #IWQNS

Patrick Thiran: Previous done in simplified boolean model. A node is either connected or not. He now refines the model to take interference into account, computing a SNR between nodes i and j. Nodes directly connected iff min{SNR_i→, SNR_j→i}>β. #LTQI #IWQNS

Patrick Thiran: Interference can destroy connectivity, but not always. The model is quite different from Poisson, with an upper bound on node connectivity. Competing effect between interference and density.
#LTQI #IWQNS

PAtrick Thiran: What is the throughput achievable per node Θ(n).
Upper bound Θ(n)=O(1/√n)
Lowerbound Θ(n)=Ω(1/log n) (?)
#IWQNS #LTQI

@threadreaderapp please kindly unroll…