2. • Anonymizing Peer-to-Peer(P2P) systems is essential to mask the
identities of the users for the privacy consideration.
• They are mainly of 2 types
─ Path based
─ Non-path based
• Rumor Riding is an example for non-path –based mutual
anonymity protocol for decentralized P2P systems.
• RR uses the symmetric cryptographic algorithms.
2
3. • Fully utilizes the resource of the whole system.
• Peers are both clients and servers in an overlay network.
• There are mainly 2 types of P2P systems
Structured P2P systems
Unstructured P2P systems
• In structured P2P networks, peers are organized following
specific criteria and algorithms, which lead to overlays with
specific topologies and properties.
• Unstructured P2P networks do not impose any structure on
the overlay networks.
3
4. • Unstructured P2P systems are of 3 types:
Centralized
Decentralized
Hybrid
5. • Anonymity is the state of being indistinguishable from other
members of some group.
• In terms of the computer community, the anonymity requirement
is especially important for those users who want to protect their
personal, private, and sensitive information, such as the user
name, ID, and IP address, during communication with others.
• Not trying to protect content of message.
• Anonymity can be
─ Path-based
─ Non-path-based
6. • Anonymous paths has to be pre-constructed.
• Uses RSA algorithm.
• Both the peer collection and content encryption introduce
high costs.
7. • No need to collect public keys for pre-construct a “secured
path”.
• Changing delivery paths often.
• Eliminating path maintenance overhead.
• Used AES cryptographic algorithm.
• Reduce cryptographic overhead for the Initiator, Responder
and Middle nodes
8.
9. • Rumor Riding is a lightweight and non-path-based mutual
anonymity protocol for P2P systems.
• Rumor Riding includes five major components:
1. Rumor Generation and Recovery.
2. Query Issuance.
3. Query Response.
4. Query Confirm.
5. File Delivery.
12. sower sa
Responder
Initiator I
Flooding Key rumor
Cipher rumor
13.
14.
15. sower sa
Responder
IPsa
Initiator I sower sb IPs Re
b
Response cipher Response key
rumor TCP Link rumor
Reversed path of
cipher rumor Reversed Path of key
rumor
16.
17. sower sc
Responder
sower sb
Initiator I
Reversed path of response TCP Link Reversed path of response
cipher rumor key rumor
Confirm cipher Confirm key
rumor rumor
18.
19. sower sc
Responder
Initiator I sower sd
Data rumor TCP Link Data rumor
20. Multiple Rumor Riding
•In RR, an initiator can issue multiple rumors in the query cycle. This
scheme is denoted as (i, j)-RR, which issues i cipher rumors and j key
rumors.
•Advantage of the multiple rumors is that RR can be more reliable as
more sowers can serve the query.
21. Rumor TTL
•The adaptive TTL determination of RR has two phases:
(a) setting initial TTL value.
(b) adaptively adjusting TTL.
Rumor Cache
•Storage overhead is related to the speed of query generation.
Ex: Each peer normally issues no more than 0.3 queries per minute
on average.
•FIFO schemes are used to handle cache overflow.
22. 1. Sower Distribution and Collision Rate
• P2P systems mainly utilize three communication patterns to
deliver messages: flooding, random walk, and end-to-end
delivery.
• Collision distance may be defined as the number of hops present
in the shortest path between the initiator and the sower.
• If this collision distance is sufficiently large then receivers are
randomly distributed in the network.
26. 2. Attacks
•Suppose network has m adversary nodes then probability of peer being
an adversary is m/n.
Collaborating attack Withstands attacks
Timing attack Withstands attacks
Traceback attack Withstands attack
Predecessor attack Withstands attacks
Traffic analysis attack Withstands attacks
28. 1.Collision rate
To verify the theoretical results . we examine the distribution of
collision rate with real traces. Besides the verification, we also use the
results to guide the selection of rumor parameters.
2.Collision distance
A longer collision distance often means a higher anonymity level,
but also increases the delay of a query as well as the traffic overhead.
On the other hand, the collision distance must be sufficiently large to
guarantee sower diversity.
3.Sower diversity
The metric reflects the distribution of sower locations in the P2P
systems. Evenly random distribution of sower location leads to a higher
anonymity degree.
29. 4.Number of sowers
Since each sower implements a selective flooding search for an
initiator, too many sowers will incur a large number of replicated query
messages, and too few sowers will result in failure on providing enough
redundancy and reliability.
5.Traffic overhead
The amount of traffic overhead represents the comprehensive
latency in data delivery and bandwidth
6.Response time
In P2P systems, it is defined as the time elapsed from when a query
is issued to when the first response arrives.
7.Crypto latency
The overhead incurred by the main cryptographic algorithms
30. • Employing a random walk concept, RR issues key rumors and cipher
rumors separately, and expects that they meet in some random peers.
• Various attacks defended by RR is also considered.
• This protocol can be even extended in other distributed systems, such
as grid systems and ad-hoc networks.
31. REFERENCES
[1] Rumor Riding: Anonymizing Unstructured Peer-to-Peer Systems Yunhao Liu, Senior
Member, IEEE, Jinsong Han, Member, IEEE, and Jilong Wang, Member, IEEE, VOL. 22,
NO. 3, MARCH 2011.
[2] Low-Cost and Reliable Mutual Anonymity Protocols in Peer-to-Peer Networks Li
Xiao, Member, IEEE, ZhichenXu, Member, IEEE, and Xiaodong Zhang, Senior Member,
IEEE, VOL. 14, NO. 9, SEPTEMBER 2003.
[3] Rumor Riding: A Protocol for Providing Anonymity Harish A, Veerapandian N
International Journal of Soft Computing and Engineering (IJSCE) ISSN: 2231-2307,
Volume-1, Issue-ETIC2011, January 2012.
[4] JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY 23(4): 660{671 July 2008
Survey on Anonymity in Unstructured Peer-to-Peer Systems Ren-Yi Xiao.
[5] http://en.wikipedia.org/wiki/Peer-to-peer.
Let ’ s first take a look at what is the Peer-to-peer model. Peer to peer mode is aiming to fully utilize the resource of the whole system. basically, peer are both clients and servers, and build an overlay p2p network. Our research will focus on the unstructured P2P architecture. the peers in the unstructured p2p systems are interconnected in an adhoc way. The unstructured p2p systems can be divided into 3 categories: the centralized, decentralized and hybrid. Centralize unstructured p2p uses a central server to perform the query service. Each node uploads the resource list, and sends query to the server, the server give them the ip address of responders, and peer download the target from a desired responder. Single point failure Decentralize unstructured p2p get rid of central server. User can use a so `call route search and direct download machnism to retrieve the object. I will talk it later. Hybrid p2p systems combine these two models. Some super peer interconnect each other in a decentralized way, and a super node and a group of normal nodes connected in a centralized way. Thus, the hybrid model improves the system efficiency.