Transcript Slide 1
Freenet
By Yogesh Kalyani
Company LOGO
Introducing P2P Overview of Freenet Key’s Protocol Security Analysis Improvement Strength and Weakness Conclusion
OUTLINE
What is Peer-to-Peer (P2P)?
“…an entity with capabilities similar to other entities in the system.”
Every end host known as peers,
and capabilities has equal roles They can act as both client and server, and provide for direct communication between them
Client Server Model Vs P2P Model
Client Server Model: Contact a server and download a web page Server has all the resources and capabilities P2P Model: Direct communication between the peers Peer acts as both client and server
Freenet
What is Freenet ?
A decentralized distributed file storage system
How does it work?
Files stored and replicated across a distributed network environment, with a peer-to-peer query and data access system. No centralized system management
Motivation
What does it provide
?
Anonymity for both producers and consumers of information Deniability for storers of information Efficient dynamic storage and routing of information Decentralization of all network functions
Key’s
Files in Freenet are identified by binary file keys, which are obtained by applying a hash function Freenet use SHA-1 function for hashing Freenet Uses three types of file keys 1) Keyword-signed key (KSK) 2) Signed-subspace key (SSK) 3) Content Hash Key (CHK)
Structure of Keys
Each Freenet key has the following structure “freenet:” is the standard prefix First three chars state key type: KSK, SSK, CHK “@” symbol separates the key type from the rest of the message Then a long set of characters used to identify the file Example:freenet:KSK@papers/p2p/freenet/keys
Keyword-signed key (KSK)
Easiest to use of all the key types KSK is derived from a short descriptive text string Ex: text/philosophy/sun-tzu/art-of-war This string is then used as input to generate a public/private key pair Public key is then hashed to yield the file key Private key is used to sign the file
Keyword-signed key
Advantages
1) Only the file description needs to be published 2) Easy to remember
Disadvantages
1) Global Namespace 2) Nothing prevents two users from independently choosing the same descriptive string for different files
Keyword-signed key
Disadvantages(2)
3) Users can abuse the names of popular files by inserting their file with the same name 4) Users can abuse the names of popular files by inserting their file with the same name
Possible because the file description is published
5) Attacker can use dictionary attack against this sign by compiling the list of descriptive string
Signed-subspace key (SSK)
Addresses the problems suffered by KSK by allowing user to make personal name space User creates a namespace by randomly generating a public/private key pair, Which will used to identify his namespace Private Key
Only the person who posses the private key can insert files to the namespace in the network Allows others to ensure a file was posted by a certain person
Public Key
Allows users to retrieve the file from the network (with Descriptive String)
Signed-subspace key
To get the key for the subspace: First he chooses a short descriptive text string like text/philosophy/art-of-war.
1) Public key is hashed 2) The descriptive string is hashed 3) (1) XOR (2) 4) (3) is hashed 5) (4) is encrypted using the file description
Content Hash Key (CHK)
A content-hash key is derived by directly hashing the contents of the corresponding file.
It gives every file a pseudo-unique file key Files are also encrypted by a randomly-generated encryption key.
Content-hash keys are most useful in conjunction with signed-subspace keys using an
indirection mechanism
For Retrieval user has to publishes the content-hash key with the decryption key
Clustering of Keys
When a node successfully receives a file from another node. It associates that node in its routing table with the hash key of the file All future requests from this node will send the request to the node listed in the routing table associated with the key closest to the key of the file being requested
Freenet: Routing
Request Hash key for data (160-bit SHA-1) Find node with closest match Forward query to this node by specifying key and #HTL Return data, replicating along the way
Freenet: Routing, Request
Freenet: Routing
Insert:
Hash key for data (160-bit SHA-1) Initiate a query with key, #HTL Node will check into its data store whether they have the data If then will response with the preexisting file assuming request has been made Else will forward the query to next node
Freenet: Routing, Insert
Process goes on till #HTL Expire If no collision i.e. or request is successful return “AllClear” Message Push data onto all the nodes (#HTL)
Freenet: Routing, Insert
Quality of routing improves over time Each time a requestor successfully receives a file from another node It adds that node to it’s routing table Associates the file key with that node in the routing table All future requests for files with similar keys will be sent to nodes associated with these keys
Freenet: Routing, Insert
Eventually a node will become specialize in serving the request for a families of keys which are close together and will store more files
Reasons:
Other node will send request for files that have similar keys If it does not have the file it forwards the request to another node based on it’s routing table When the file is located, the response gets passed back Each node on the responses path gets a copy of the file stored This includes the node the request was initially sent to Over time this node will start to store more and more files with this key type
Freenet: Routing
Managing Data
Node storage is managed as an LRU (Least Recently Used) cache in which data items are kept sorted in decreasing order by time of most recent request or time of insert, if an item has never been requested
Freenet: Routing, Managing Data
Each node has limited storage capacity The user has no knowledge of what files their node stores As well No user knows the identity of a node that provides a file they have requested Or knows the identity of a node that has requested a file from them
Freenet: Routing, Managing Data
NO file lifetime guarantees Popular files will spread to many nodes Each requested file located, will be copied to every node it passes through on the path from the source node to the requestor node Rarely accessed files will slowly be removed from the network as room is required for new files i.e. removing least-recently used (LRU) files As a node runs out of space, files will be deleted in order of least recently requested to make room Rarely requested files will becomes limited
ONLY
be removed if space
Protocol: Protocol Descriptors
Request.Handshake
Reply.Handshake
Initial Contact Request.Data
Send.data
Querying for data
Protocol: Protocol Descriptors
Reply.NotFound
Reply.Restart
Failure Message Request.Continue
Request.Insert
Request Management Reply.Insert Send.Insert
Inserting Data
Security Analysis
Eavesdropping
No protection against eavesdropping between the user and the first node contacted in basic Freenet Users are recommended to first connect to a node running on their own machine in basic Freenet Pre-Routing to protect from eavesdropping in new version of freenet
Security Analysis: Eavesdropping
As routing depends on knowledge of the search key Key anonymity is not possible.
Use of hashes as keys provides some obscurity against casual eavesdropping But still vulnerable to dictionary attacks Can be easily done by setting up a node and "eavesdropping" any messages that are routed through her
Denial of Service Attacks
An attacker can insert a large number of junk files into the network Counter attack: By dividing the datastore into two sections, one for new inserts and one for established files New inserts would only replace new inserts; therefore a flood of them could not displace existing files
Denial of Service Attacks
Problem with dividing Datastore:
Difficult for genuine new inserts to survive long enough to be requested by others and become established An attacker may be able to legitimize his bogus files, by requesting them from strategic locations where it will be cached on as many nodes as possible
Traffic Analysis
Traffic analysis is possible by setting up nodes in strategic locations and by monitoring encrypted traffic between nodes Achieved by observing
Search Key Closeness
, Hope-to-live, Depth.
Simplest case, if a node receives a request for a search key that it has never served anything close to, he can guess that the request originated from a directly connected node or that he is very early in the chain .
Improvement
Original Algorithm in Freenet:
The assessment of whether a peer X has common interests with a peer Y is based on historical data, i.e., on whether peer X was able to respond to previous queries sent by peer Y
Improvement
*New algorithm[3]
Principle: The "current" content stored at a peer reflects its current interest. Thus, common interest between two peers is assessed directly from the current content locally stored at both peers and is periodically re-evaluated In this way each peer maintains a community of peers, which share similar interests i.e. number of files
study shows :
It reduces the average Freenet request and response path lengths by up to 39%
Freenet: Design strength
Anonymity for both producers and consumers of information Completely Decentralized: No single point of failure Scales well Dynamic routing adapts well to changing network topology High resilience to attacks
Freenet: Design weakness
No file lifetime guarantees Not an efficient keyword search Currently, no defense against DoS attacks
Conclusion
Freenet uses depth first search with backtracking which makes it scalable when the size of network is increased or network is overloaded Freenet is a loosely unstructured decentralized network, in which every node has equal roles and do the same task thereby increasing the fault tolerance Overall, Freenet is good P2P network for file sharing system; deploying efficient routing for the location and the storage of the data
References
Freenet, http://freenet.sourceforge.net.
Freenet documentation, http://freenet.sourceforge.net/doc/book.html.
Marcelo Werneck Barbosa, “Using locality of reference to improve performance of peer-to-peer applications” Proceedings of the fourth international workshop on Software and performance, ACM Press,
Year of Publication: 2004, Pages: 216 - 227
Clarke, O. Sandberg, B. Wiley, and T. W.Hong, “Freenet: A distributed anonymous information storage and retrieval system”, in
ICSI Workshop of Design Issues in Anonymity and Unobservability,
July 2000.
SM Lui, SH Kwok, “ Interoperability of Peer-To-Peer File Sharing Protocols ” ACM SIGecom Exchanges, Volume 3, Issue 3 Summer, 2002, ACM Press, Year of Publication: 2002 Pages: 25 - 33