It is just not any network; it is an anonymity network whose prototype has been developed by Massachusetts Institute of Technology (MIT) and the École Polytechnique Fédérale de Lausanne (EPFL). It is promised to be both bandwidth and computation efficient which will allow clients from low bandwidth to connect and will allow servers to serve more clients. One problem that exists in Tor is that a malicious user can flood a node in the TOR network with own crafted messages destined to a predefined node and observe the path of a single message which is to be tracked. Even though the attacker cannot see what’s inside the message since it is encrypted but definitely path trace the path of the single message. Riffle seems to have addressed this problem. At a very basic, Riffle consists of clients and a distributed system of nodes as has been the case always with these types of anonymous network models. Instead of sending a message to a single server, Riffle will send a message to all servers where messages shuffle will happen. For example, if five people send a message then they will be reshuffled each time it will reach the successive node. In real world scenario case, there will be thousands of messages floating along the nodes of Riffle network all mixed up, and thus it will be nearly impossible to predict the path of the message. Riffle uses Mix Networks (mixnet) whose aim to collect user inputs and mixes them before sending out. However, if there is any malicious mix, then the mixnet will fail to provide the level of anonymity that the model states. To overcome the malicious mixnet issue, Riffle uses a technique called verifiable shuffle, and it works on top of the Onion protocol. In the verifiable shuffle, when a mix shuffles the inputs it will also generate a mathematical proof. It all starts when a client wants to achieve anonymity will send an initial message to all nodes in the Riffle network to establish a mathematical proof. The generated proof can be used by other parties along with ciphertext to make sure that the mix has not been tampered with. Now this verification sounds ok but considers the performance implications this model will have to bear in case of thousands of messages. But the Riffle has already addressed this problem with the workaround of Authentication Encryption. In Authentication Encryption, whenever a client initiates a message, a unique private key will be generated to encrypt the message and will be used by the nodes in the network. This key will be sent by clients to all nodes, and then the subsequent operation of verifying the message at each node will be done. This gives us a better protection that even if a single server is honest in the network, then RIFFLE is secure.
Riffle is more secure than TOR as it relies on an anytrust model. Therefore, as long as a single server in the anonymity network is safe, Riffle is secure. Tor is criticized heavily for its poor performance. Riffle is faster than TOR. In the research, it was found out that file transfers in Riffle took one-tenth of the time required for TOR and other anonymity networks. In file sharing Riffle can achieve a bandwidth of over 100KB/per client in an anonymous network of 200 clients.
All major details of Riffle Anonymous Network can be found in the referenced link.
