Summary by Sasha

- the fundamental idea behind this paper is in providing a P2P application
- that will support anonymity of both authors and readers

- protocol does not guarantee data availability or life-time; if data is not
- popular it will fade out and eventually all cached copies of data will
- disappear

- protocol learns usage patterns through node requests and responses and
- builds most optimal routes

- routes are based on semantic-closeness of data which is built not on
- data/file descriptions (which are textual) but on hashes, which evenly
- distribute any related files over the network

- author discusses different tasks in the network; retrieving data, storing
- data

- naming together with searching and updating is a serious part of research
- involving additional approaches, like additional indirect (file) access,
- signature

-verified indirect files, contenthash keys, etc

- they discuss splitting files in smaller chunks with references stored in
- indirect files

- further, author discusses security issues, possible attacks both from
- internal and external attacks, presenting importance of non-probabilistic
- routing that protects against statistical analysis attacks

Summary by Adnan

Single point of failure is an important aspect under data management.  For
instance if we take it from security point of view then, if a centralized
system is compromised then it can cause problems at a bigger scale.  A huge
number of audiences will be affected. There are various solutions out there
dealing with data management based on either centralized or decentralized
system. This paper introduces such a system which is based on the concept of
decentralized system. Free net is a system which allows to store data over a
network on different nodes. The paper does a detailed research on how the data
will be stored, fetched on/from the nodes.  The data is stored among different
nodes, so the nodes availability is important aspect. Freenet is software
available to download. Freenet is based on five main design goals including
Anonymity for both producers and consumers of information, Deniability for
storers of information, Resistance to attempts by third parties to deny access
to information, Efficient dynamic storage and routing of information and
Decentralization of all network functions.  Freenet is implemented as a peer
to peer network of nodes that query one another to store and retrieve data
les which are named by location independent keys. A packet in a freenet
protocol contains self-oriented messages. It supports both TCP and UDP but it
prefers TCP, when available. Naming, searching and updating data has been
discussed in the paper and concept of content hash key is used for that.
Performance has been measured by carrying out simulation. Different security
aspects are also considered specially routing. The Freenet network provides an
effective means of anonymous information storage and retrieval. The work on
freenet is still in progress bit around 15,000 copies has been download and
it¨ªs in use.

Summary by Patrick

In this paper the authors propose a peer-to-peer distributed storage system
that allows  anonymous storing, replication and retrival of files while
protecting the identity of both publishers and readers.

Files are stored using either one of two kinds of keys. Content-Hash key and
signature-verifying key. The content-hash key is a hash based on the content
of the file (SHA-256) and the signature-verifying key is based on
public/private keys.

Evaluation is done by simulating a system and comparing the successful
requests against a different number of nodes. I found that the evaluation part
was lacking in details and they could probably have done a more thorough job.

The authors also talk about different security aspects, how to protect
anonymity, prevention of insert-spamming, modification of files and other
potential attacks. However, I don't think they cover a lot of potential attack
vectors. They suggest using a hash cash scheme to protect against users from
spamming large garbage files in order to overwrite genuine files, but it would
be possible to attack such a system with a botnet.

An attack vector not covered by the paper, but that I think is quite important
is that encryption is optional. They suggest that files should be encrypted to
protect the storage-users from knowing what they are storing. A malicious user
(or a group of users) could spam the network with unencrypted illegal files to
threaten both the networks authenticity as well as creating fear for
individual storage users.

Summary by Zahid

Freenet: A Distributed Anonymous Information Storage and Retrieval System

Summary:

The computer networks are growing very rapidly in a large scale so the
importance of storage and exchange of information is also going high.
Regarding to user¡¯s privacy  the information is typically store only one or a
few fixed places so the privacy is on high risk. There are many reasons but on
of them is to creating a central point to store or exchange the information.
It would be very easy for opponents to remove or hack the data from a central
system or location.

System need greater security and reliability, just because of this reasons the
authors suggest new system called Freenet, a distributed information storage
and retrieval system design to address these concerns of privacy and
availability.  Freenet works under the concept of peer-to-peer network of
nodes that query one another to store and retrieve data files, which are named
by location independent keys.

Freenet has one good future that, it enables users to share unused disk space.
Freenet operates as a location independent distributed file system across many
individual computers that allows files to be inserted, stored and requested
anonymously.

Freenet has five main goals like as follow

1) Decentralization of all network functions. 2) Efficient dynamic storage and
routing of information. 3) Anonymity for both producers and consumers of
information 4) Resistance to attempts by third parties to deny access to
information 5) Deniability for storers of information

According to authors that it is not intended to guarantee permanent file
storage, although it is hoped that enough nodes will join with enough storage
capacity that most tiles will be able to remain indefinitely. Freenet is a
free software project on Sourceforge.

Summary by Ioana

Freenet: a distributed anonymous information storage and retrieval system.
	Ian Clarke, Oskar Sandberg, Brandon Wiley, and Theodore W. Hong. 

In this paper, the authors propose Freenet, a peer-to-peer application that is
designed to offer anonymity of sender and receiver of the data. The system is
constructed of a network of identical nodes, which provide storage for data
files and cooperate to obtain efficiency routing of messages.

The proposed solution is based on queries send by the nodes in the network to
retrieve or insert data files. The messages are passed from a node to another,
and a node knows only the previous and the next hop.

There are two types of transactions that can occur : retrieve and insert data
in the networks. For each of these two, different messages are used :

- Retrieve data : ¡°Request.data¡± send by the node that is requesting the data; 
		  ¡°Reply.Restart¡± send by remote nodes, while the request is processed;
		  ¡°Send.Data¡±, if the request is successful 
		  ¡°Reply.NotFound¡± in case the data which was requested is not found.
- Insert data :   ¡°Request.Insert¡±
		  ¡°Reply.Restart¡± ( same as for retrieving data)
		  ¡°Send.Data¡± and ¡°Reply.NotFound¡±
		  ¡°Request.Continue"
		  "Reply.Insert" and ¡°Send.Insert¡±;

Apart from the messages exchanged during the two transaction, in the beginning
of such an action a ¡°Request.Handshake¡± and ¡°Reply.Handshake¡± are exchanged.

In order to evaluate the performance of the system, the authors use
simulations in which different number of nodes are used( from 500 to 900
nodes), each having a storage size of 40 items and a routing table of 50
addresses. The authors evaluate the retrieval time and the retrieval success
rate. The main findings are : - for the retrieval time, ( measured in number
of hops needed for a successful request):  for all the size of the networks,
the value stabilises around  10 request. - for the retrieval success rate :
the number of queries necessary until the network converges is approximately
half of the network size( for 500 nodes this value is equal to 300, whereas
for 900 nodes is less than 600).

In order to evaluate the security, the authors employ a taxonomy of anonymous
communication properties on three axes : sender anonymity/receiver anonymity ,
local eavesdropper and the degree of anonymity.