# Outsourced Multiparty computation proof in Ideal world

I need to know in an outsourced two party computation where honst $A$ and $B$ outsource their private and secure data to a malicious server, why we need to design a simulator that interacts with an adversary who exists in real model?

See the paper cited below as an example:

https://eprint.iacr.org/2011/272.pdf

• you need to design the simulator in order to prove security. See this answer – mikeazo May 13 '15 at 12:03
• The start of section 4 of the paper you linked to has a pretty good explanation. – mikeazo May 13 '15 at 12:04