Providing Witness Anonymity in Peer-to-Peer Systems

CSCE 715
Ankur Jain
11/16/2010
 Introduction
 Design Goals
 Framework
 SDT Protocol
 Achievements of Goals
 Overhead of SDT
 Conclusion
 Peer-to-Peer systems
 Distributed application architecture
 Partitions task between peers equivalently.
 E.g. – Skype, Cloud Computing, P2PTV and
many more.
 Fundamental Challenge
 Trust relationship between peers.
 Several research studies.
 To build trust and reputation between peers.
 Reliability
 Computing true trust value.
 Presence of malicious user.
 Anonymity
 Non Identification of peers
 Accountability
 Identification of malicious peers.
 Previous research focused on reliability.
 Overall Goal
 Extend P2P trust management systems
 To provide Witness Anonymity
 To provide anonymity to person reporting
malicious behavior.
 To preserve privacy of peers.
 To hide trust topology from malicious parties.
 Identity Anonymity
 Backward Anonymity
 Traceability
 Non-slanderability.
 Additional Goals
 Efficiency
 Decentralization.
 System Model
 No Trusted Third Party.
 2 types of user
 Offline Group Manager (OGM)
 User
 # of adversaries less than threshold t.
 Adversary Model
 2 types of adversaries
 Malicious user
 Selfish user
 Will collude together to maximize the attack.
 Network Model
 Mixnet based anonymous system
 Consist of series of servers called MIXes.
 Associated with public keys.
 Receives encrypted messages.
 Decrypts, batches, permutes, forwards messages.
 Strips off sender’s name and identifying information.
 Mechanism for monitoring claims sent
 Irrespective of claims being generated or forwarded.
 SDT – Secure Deep Throat
 Provide anonymity and accountability together.
 Include tracing mechanism to identify user.
 4 step procedure
 Setup
 Registration
 Claim Broadcasting
 Public Tracing
 Modes of Operation
 Active: Real Time requirements.
 Passive: Not strict Real Time requirements.
 OGM generates public and secret keys.
 Identification list (LIST) initially empty.
 Define tag bases
 Used in claim broadcasting
 To create anonymous claims.
 Only one per type of misbehavior per user.
 User contacts OGM.
 User selects identity.
 Check its availability.
 User obtains a member public/secret key pair.
 OGM adds a new entry to LIST.
 OGM select s items from LIST and sends it to user.
 User sends confirmation for key pair and LIST items received.
 User maintain two databases.
 Maintains claim sent by herself.
 Maintains claim received from other user.
 On detecting malicious behavior
 Checks database for previous entries for same type of
behavior.
 If not found generates new claim using tag base.
 Broadcast through anonymous communication
system.
 Also stores claim in database.
 On receiving claim
 Checks whether entry for that claim is present or
not.
 If yes, then drops the claim.
 If not check its validity and stores the claim.
 Also forwards it in the system.
 Initializing Public Tracing
 User finds t claims.
 Checks distinctness of all t claims.
 Generates a message including t claims and
broadcast it to network
 Check for entries in databases.
 If found broadcast two entries as proof to disclose
the identity of malicious user.
 If no entries found broadcast message NO-ONE.
 After receiving NO-ONE message other repeat the
steps in their local LIST.
 Used when real time requirement is not critical.
 Achieve better efficiency.
 Changes in claims broadcasting
 Claims regarding malicious behavior not sent
immediately.
 Sent these claims only when queried about the
behavior of user.
 Public tracing will performed on all claims to
prevent multiple claims from an adversary.
 Peer forwards claim with a probability.
 Instead of flooding entire network.
 Lower the probability, lower is the number of peers
storing the claim.
 Lower is the probability that one peer stores every t
distinct claims.
 Require more number of witnesses in this case.
 Also non zero probability that adversary may escape
disclosure.
 Identity Anonymity
 May be broken using Traffic Analysis or Protocol
Analysis
 Traffic Analysis is prevented by Mixnet based
communication system.
 Protocol Analysis is also hard to perform
 No public key in claim broadcasted
 All parameter are calculated using discrete
algorithm so very robust against brute force attack.
 Backward Anonymity
 Adversaries can compromise multiple peers.
 Claim does not provide information regarding identity.
 No way to differentiate the user on basis of claims.
 Also ensured when OGM and adversaries are in contact
 User’s secret key is only known to user.
 No way to extract secret key from OGM.
 Traceability
 Good peers need to find a valid record of adversary
from LIST.
 LIST items are distributed among different peers.
 Probability of all copies controlled by adversary
group is very small.
 Non-Slanderability
 Max number of claims sent by adversaries against a
user
 Total number of adversaries which is less than t.
 Adversaries cannot collect enough claims to
remove good user from the system.
 Distributed storage of LIST
 OGM maintains LIST offline.
 LIST is stored in distributed form.
 Peers do not have knowledge of LIST items with
other peers.
 Helps in detecting a adversary even if adversary is
controlling the majority of LIST.
 Communication Costs
 Major cost is forwarding claims.
 Implemented using elliptic curve or hyper elliptic
curve over a finite field.
 Claim size not more than 409 bytes.
 LIST distribution another cost.
 Smaller the LIST, higher probability of message
broadcast while tracing.
 Storage Requirements
 For cryptographic keys, LIST and local databases.
 Storing personal keys and public key of OGM.
 Only small part of the entire LIST.
 Very small database requirement in passive mode.
 A probabilistic forwarding approach may reduce
database space in active mode.
 SDT provide witness anonymity to users reporting
malicious behavior.
 Two modes of operation: Active and Passive.
 Overhead is acceptable in peer-to-peer systems.