Understand the Multi-party circuit: Priority Encoder

1 reply · 2K views · last reply
zak100
Messages
462
Reaction score
11
TL;DR
Hi, I am reading a paper from butI can't understand the circuitry
Paper link is:

MPCircuits: Optimized Circuit Generation for Secure Multi-Party Computation

I have a question:
What is the reason for using priority encode?

The circuit diagram is given. I want to give some text from book:

In secure stable matching, the match list is computed while
keeping the preference lists private to their respective owners.
This problem has been studied in the recent literature [21],
[22] where the secure stable matching problem is reduced to a
two-party secure computation scenario. Each individual XORshares
her preference list and sends it to two non-colluding
servers who perform the secure computation. However, stable
matching is inherently a multi-party problem and the assumption
of two non-colluding servers may not be feasible in
practice. To the best of our knowledge, we provide the first
solution for multi-party secure stable matching.
From the above text I want to understand why its a multi-party? It is taking input of 2 groups but why this circuit works as ##n * n## and not as ##2 * 2##?

Somebody please guide me
Block diagram of a stable matching circuit for round r_j, showing inputs from the previous round, a Priority Encoder, preference list tables from group 1 and group 2, multiplexers, an update logic block and outputs to the next round.

Zulfi.
 
Attachments
  • Block diagram of a stable matching circuit unrolled for round r_j, showing inputs is_free_group1, is_free_group2, C[j] and match_list, a priority encoder, preference lists L0 to Ln/2 and Ln/2+1 to Ln, muxes, a C[g] lookup, a more preferred comparator and update logic.
    How we have multiparty stable matching.jpg
    46.4 KB · Views: 359
Last edited:
Hi,
The quotation is not displaying some text.

Zulfi.