SLIDE 1
Secure Multi-party Computation What it is, and why youd care Manoj - - PowerPoint PPT Presentation
Secure Multi-party Computation What it is, and why youd care Manoj - - PowerPoint PPT Presentation
Secure Multi-party Computation What it is, and why youd care Manoj Prabhakaran University of Illinois, Urbana-Champaign SMC SMC SMC conceived more than 30 years back SMC SMC conceived more than 30 years back A very general concept that
SLIDE 2
SLIDE 3
SMC
SMC conceived more than 30 years back
SLIDE 4
SMC
SMC conceived more than 30 years back A very general concept that subsumes the bulk of theoretical cryptography
SLIDE 5
SMC
SMC conceived more than 30 years back A very general concept that subsumes the bulk of theoretical cryptography Largely a well-kept secret
SLIDE 6
SMC: the question
SLIDE 7
SMC: the question
Collaboration without trust?
SLIDE 8
SMC: the question
Collaboration without trust? Collaboration: compute on collective data belonging to different parties
SLIDE 9
SMC: the question
Collaboration without trust? Collaboration: compute on collective data belonging to different parties e.g. query with me, database with you
SLIDE 10
SMC: the question
Collaboration without trust? Collaboration: compute on collective data belonging to different parties e.g. query with me, database with you e.g. query with me, encrypted database with you, key with someone else
SLIDE 11
SMC: the question
Collaboration without trust? Collaboration: compute on collective data belonging to different parties e.g. query with me, database with you e.g. query with me, encrypted database with you, key with someone else Goal: Nothing should be revealed “beyond the result”
SLIDE 12
SMC: the question
Collaboration without trust? Collaboration: compute on collective data belonging to different parties e.g. query with me, database with you e.g. query with me, encrypted database with you, key with someone else Goal: Nothing should be revealed “beyond the result” “Ideally”: Use a trusted third party
SLIDE 13
SMC: the question
Collaboration without trust? Collaboration: compute on collective data belonging to different parties e.g. query with me, database with you e.g. query with me, encrypted database with you, key with someone else Goal: Nothing should be revealed “beyond the result” “Ideally”: Use a trusted third party “Really”: Can’t agree on a trusted party. So...
SLIDE 14
SMC: the answer
SLIDE 15
SMC: the answer
SMC protocol: among mutually distrusting parties, to emulate the presence of a globally trusted party
SLIDE 16
SMC: the answer
SMC protocol: among mutually distrusting parties, to emulate the presence of a globally trusted party Numerous protocols in literature for various functionalities, in various settings
SLIDE 17
SMC: the answer
SMC protocol: among mutually distrusting parties, to emulate the presence of a globally trusted party Numerous protocols in literature for various functionalities, in various settings Tools: Verifiable secret-sharing, homomorphic encryptions, commitments, ZK proofs, oblivious transfer, ...
SLIDE 18
SMC: the answer
SMC protocol: among mutually distrusting parties, to emulate the presence of a globally trusted party Numerous protocols in literature for various functionalities, in various settings Tools: Verifiable secret-sharing, homomorphic encryptions, commitments, ZK proofs, oblivious transfer, ... Simpler protocols if some trust already present
SLIDE 19
SMC: the answer
SMC protocol: among mutually distrusting parties, to emulate the presence of a globally trusted party Numerous protocols in literature for various functionalities, in various settings Tools: Verifiable secret-sharing, homomorphic encryptions, commitments, ZK proofs, oblivious transfer, ... Simpler protocols if some trust already present “Honest-but-curious”
SLIDE 20
SMC: the answer
SMC protocol: among mutually distrusting parties, to emulate the presence of a globally trusted party Numerous protocols in literature for various functionalities, in various settings Tools: Verifiable secret-sharing, homomorphic encryptions, commitments, ZK proofs, oblivious transfer, ... Simpler protocols if some trust already present “Honest-but-curious” “Honest-majority”
SLIDE 21
SMC: the answer
SMC protocol: among mutually distrusting parties, to emulate the presence of a globally trusted party Numerous protocols in literature for various functionalities, in various settings Tools: Verifiable secret-sharing, homomorphic encryptions, commitments, ZK proofs, oblivious transfer, ... Simpler protocols if some trust already present “Honest-but-curious” “Honest-majority” Simple (offline) trusted sources
SLIDE 22
SMC in GENI?
SLIDE 23
SMC in GENI?
Where privacy is needed
SLIDE 24
SMC in GENI?
Where privacy is needed e.g. Measurement archives held by a virtual trusted party
SLIDE 25
SMC in GENI?
Where privacy is needed e.g. Measurement archives held by a virtual trusted party Secure distributed storage and computation (secure unless all servers corrupt)
SLIDE 26
SMC in GENI?
Where privacy is needed e.g. Measurement archives held by a virtual trusted party Secure distributed storage and computation (secure unless all servers corrupt) May use “honest majority” in a federation
SLIDE 27
SMC in GENI?
Where privacy is needed e.g. Measurement archives held by a virtual trusted party Secure distributed storage and computation (secure unless all servers corrupt) May use “honest majority” in a federation Provide SMC as an “experiment support service”?
SLIDE 28