Citation Context

...atency. The anonymity protocol that offers the strongest guarantees is the DiningCryptographers protocol, also known as Dining-Cryptographers network or short DC-net, which was introduced by Chaum in =-=[3]-=-. The protocol provides unconditional communication anonymity for senders and recipients, however, at the cost of low throughput and high latency, in particular when scaling to many participants. Dini...

Citation Context

...proach proposed by Chaum in [3] is to use a contention algorithm with discrete time slots and resolve collisions by retransmitting the messages. A common example of such an algorithm is slotted ALOHA =-=[14]-=-. In Slotted ALOHA participants pick a time slot for transmission at will. Whenever a collision happens, participants wait for a random amount of time before they pick a new time slot for retransmissi...

Citation Context

...d node, the proxy, is compromised. Stronger guarantees are offered by a cascade of proxies in onion-routing networks like Tor [17], which were originally proposed by Syverson, Goldschlag, and Reed in =-=[16]-=-. However, onion-routing networks are possibly susceptible to attacks that correlate traffic entering and leaving the network. See, for example, [11]. Mix-nets, proposed already in 1981 by Chaum in [4...

Citation Context

...ally proposed by Syverson, Goldschlag, and Reed in [16]. However, onion-routing networks are possibly susceptible to attacks that correlate traffic entering and leaving the network. See, for example, =-=[11]-=-. Mix-nets, proposed already in 1981 by Chaum in [4] do not have this problem; however, they suffer from increased latency. The anonymity protocol that offers the strongest guarantees is the DiningCry...

Citation Context

... until a collision has been resolved. In [13, Sec. 3.1.2.3.2], A. Pfitzmann presents such an algorithm, which is an improvement of the tree algorithm that was independently proposed by Capetanakis in =-=[2]-=- and by Tsybakov and Mikhailov in [18]. Pfitzmann calls this algorithm tree-like collision resolution with superposed receiving ; in the following, we refer to this algorithm as Pfitzmann’s algorithm....

Citation Context

... In [13, Sec. 3.1.2.3.2], A. Pfitzmann presents such an algorithm, which is an improvement of the tree algorithm that was independently proposed by Capetanakis in [2] and by Tsybakov and Mikhailov in =-=[18]-=-. Pfitzmann calls this algorithm tree-like collision resolution with superposed receiving ; in the following, we refer to this algorithm as Pfitzmann’s algorithm. Let the number of messages that colli...

Citation Context

... messages transmitted through DC-net as in the original Pfitzmann’s algorithm. The first protective measure against disruption during the scheduling phase was presented by Waidner and B. Pfitzmann in =-=[20,21]-=-. The idea is to investigate the reservation phase in case of impossible results of the specific scheduling algorithm used. For example, in Pfitzmann’s algorithm, the number of initially collided mess...

Citation Context

...d by output commitments. In the special phase called palaver phase any participant can start an investigation of the scheduling phase in order to detect disrupters. Secure multi-party computation. In =-=[10]-=-, Golle and Juels propose two new versions of DC-net that allow detection and identification of disrupters with high probability by using zero-knowledge proofs. They do not consider the problem of col...

Citation Context

...t a larger message wait. As soon as only two messages collided, only one participant retransmits; the other message is recomputed locally. A more detailed description of the algorithm can be found in =-=[19]-=- and [7, Sec 3.2.2]. Another algorithm, using a similar approach, was presented in [1] by Bos and Boer. It also computes a retransmission schedule for collided messages and requires s retransmissions ...

Citation Context

...nd verifiable DC-net, depending on the presence of disruption. For scheduling, Verdict uses a similar approach as [15,7] and the same as in another implementation of DC-net by the same group, Dissent =-=[5,22]-=-. 6 Anna Krasnova, Moritz Neikes, and Peter Schwabe In [8], Franck proposes a scheduling for DC-net based on the collision-resolution protocol SICTA. This scheduling protocol is very similar to Pfitzm...

Citation Context

...transmits; the other message is recomputed locally. A more detailed description of the algorithm can be found in [19] and [7, Sec 3.2.2]. Another algorithm, using a similar approach, was presented in =-=[1]-=- by Bos and Boer. It also computes a retransmission schedule for collided messages and requires s retransmissions after s messages collided. However, it has a larger overhead in header messages and is...

Citation Context

...nd verifiable DC-net, depending on the presence of disruption. For scheduling, Verdict uses a similar approach as [15,7] and the same as in another implementation of DC-net by the same group, Dissent =-=[5,22]-=-. 6 Anna Krasnova, Moritz Neikes, and Peter Schwabe In [8], Franck proposes a scheduling for DC-net based on the collision-resolution protocol SICTA. This scheduling protocol is very similar to Pfitzm...

Citation Context

...the message (See Appendix A for details). The length of the scheduling message can be reduced if instead of bits representing rounds, one would use elements of the additive group of integers modulo m =-=[12]-=-. After all the scheduling messages of participants are added up, elements of value 0 indicate an unreserved round, elements of value 1 indicate a reserved round, all other values indicate collisions....

Citation Context

...e position of the request in the vector. This idea is used in the Master’s thesis of Franck [7], in which he derives a fully verifiable variant of DC-net. Later, verifiable DC-net was rediscovered in =-=[6]-=- and implemented under the name Verdict. The advantage of Verdict is that it allows switching between traditional DC-net and verifiable DC-net, depending on the presence of disruption. For scheduling,...

Citation Context

...be quadratic in the number of participants to avoid collisions with high probability. Reservation maps are used by Herbivore, an implementation of DCnet presented by Goel, Robson, Polte, and Sirer in =-=[9]-=-. Herbivore optimizes the size of the scheduling message by allowing some collisions during message cycle depending on the message size (collisions for smaller messages are more likely). For performan...

Citation Context

...ze his own request only after the permutation is completed. His reserved round number can be derived from the position of the request in the vector. This idea is used in the Master’s thesis of Franck =-=[7]-=-, in which he derives a fully verifiable variant of DC-net. Later, verifiable DC-net was rediscovered in [6] and implemented under the name Verdict. The advantage of Verdict is that it allows switchin...

Citation Context

.... For scheduling, Verdict uses a similar approach as [15,7] and the same as in another implementation of DC-net by the same group, Dissent [5,22]. 6 Anna Krasnova, Moritz Neikes, and Peter Schwabe In =-=[8]-=-, Franck proposes a scheduling for DC-net based on the collision-resolution protocol SICTA. This scheduling protocol is very similar to Pfitzmann’s collision resolution algorithm, the only difference ...

Citation Context

...he problem can be avoided through techniques like secure multi-party computation of a secretly distributed permutation of slots among players, but this is impractical”. Studholme and Blake propose in =-=[15]-=- a way to implement such a multi-party computation called secret shuffle by organizing a Mix-net with participants of DC-net serving as nodes. Encrypted round-reservation requests are transmitted thro...