Citation Context

... behavior of the LT process using first principles of probability theory, which precisely captures the behavior of the data recovery process. The “digital fountain approach” concept introduced in [7] =-=[6]-=- is similar to that of a universal erasure code, and LT codes are the first full realization of this concept. This paper describes some of the theoretical underpinnings of some portions of the work de...

Citation Context

...nsmission schemes which typically wait for a message to be lost before sending another copy. At the same time, erasure coding techniques have long been used to cope with partial data loss efficiently =-=[3]-=-. A natural question to investigate is whether erasure coding techniques can be beneficial in DTN context as well. Applying erasure coding to DTNs turns out to be an interesting and non-trivial exerci...

Citation Context

...ine of upto 8 seconds. As a consequence the size of the the packet buffer for NAK-based retransmissions is 128. The packet buffer will be larger for longer deadlines. � The Digital Fountain technique =-=[3]-=-, uses Tornado codes that require the receiver to correctly �§��������� receive packets to recover � the data packets, � where is a small constant. 10sWe performed detailed experiments with a wide-ran...

Citation Context

...et the source node split the packet into smaller fragments. Out of these fragments a number of fragments is created by applying a certain coding scheme known as erasure codes or fountain codes [124], =-=[125]-=-. These codes have the property that it suffices for the receiver to successfully receive any out of the coded fragments to successfully decode the packet. This has for example been explored in [120]....

Citation Context

...ded symbols are needed, and the k source symbols can be reconstructed with high probability. In practice, the overhead parameter ǫ can be in [0.03, 0.05], depending on the specific coding that we use =-=[16, 5]-=-. For the specific case of Push-to-Peer, to ensure that a peer can reconstruct a window of size W by downloading from y different boxes, each box has to serve a fraction of size W ∗ (1 + ǫ)/(y + 1) of...

Citation Context

...he decoder, no matter from which transmitter. The decoding algorithm must also know the seeds of all the pseudo-random number generators in all the encoders. This is reminiscent of a digital fountain =-=[40]-=- but we use real valued coefficients and strictly sparse signals. 6 Conclusions The extension of LDPC matrices from a finite alphabet setting to measurement of real-valued inputs is very promising, an...

Citation Context

...tions that implies the need for keeping minimal state in the network or servers [35, 36]. Video Coding: Layered or multiple description coding offers more graceful degradation of video quality (e.g., =-=[17]-=-). However, they impose higher overhead on content providers and the delivery infrastructure and thus we do not consider this class of solutions. 9. CONCLUSIONS With the growth of video traffic, we ar...

Citation Context

...r arbitrary network conditions. However, this simple form of FEC was developed for streaming settings, where a single sender is transmitting data at a high, steady rate such as in bulk file transfers =-=[7]-=- or in a video or audio feeds [8]. Part of our challenge is to develop a FEC solution matched to the characteristics of a MANET. 3.2 Algorithm We noted earlier that Mistral can be built on top of any ...