Analysis of multipath Routing-Part I: the effect on the packet delivery ratio

In this paper, we develop an analytical framework for evaluating multipath routing in mobile ad hoc networks. The instability of the topology (e.g., failure of links) in this type of network due to nodal mobility and changes in wireless propagation conditions makes transmission of time-sensitive information a challenging problem. To combat the inherent unreliability of these networks, we propose a routing scheme that uses multiple paths simultaneously by splitting the information between a multitude of paths, so as to increase the probability that the essential portion of the information is received at the destination without incurring excessive delay. Our scheme works by adding an overhead to each packet, which is calculated as a linear function of the original packet bits. The resulting packet (information and overhead) is fragmented into smaller blocks and distributed over the available paths. The probability of reconstructing the original information at the destination is derived in an analytical form and its behavior is studied for some special cases. It is shown that, under certain constraints, the packet dropping probability decreases as the number of used paths is increased.     

Multipath routing in the presence of frequent topological changes

In this article we propose a framework for multipath routing in mobile ad hoc networks and provide its analytical evaluation. The instability of the topology (e.g., failure of links) in these types of networks, due to nodal mobility and changes in wireless propagation conditions, makes transmission of time-sensitive information a challenging problem. To combat this inherent unreliability of these networks, we propose a routing scheme that uses multiple paths simultaneously by splitting the information among the multitude of paths, to increase the probability that the essential portion of the information is received at the destination without incurring excessive delay. Our scheme works by adding some overhead to each packet, which is calculated as a linear function of the original packet bits. The resulting packet (information and overhead) is fragmented into smaller blocks and distributed over the available paths. Our goal is, given the failure probabilities of the paths, to find the optimal way to fragment and then distribute the blocks to the paths so that the probability of reconstructing the original information at the destination is maximized. Our algorithm has low time complexity, which is crucial since the path failure characteristics vary with time and the optimal block distribution has to be recalculated in real time     

Prediction of signal peptides in archaea

Computational prediction of signal peptides (SPs) and theircleavage sites is of great importance in computational biology;however, currently there is no available method capable of predictingreliably the SPs of archaea, due to the limited amount of experimentallyverified proteins with SPs. We performed an extensive literaturesearch in order to identify archaeal proteins having experimentallyverified SP and managed to find 69 such proteins, the largestnumber ever reported. A detailed analysis of these sequencesrevealed some unique features of the SPs of archaea, such asthe unique amino acid composition of the hydrophobic regionwith a higher than expected occurrence of isoleucine, and acleavage site resembling more the sequences of gram-positiveswith almost equal amounts of alanine and valine at the position-3before the cleavage site and a dominant alanine at position-1,followed in abundance by serine and glycine. Using these proteinsas a training set, we trained a hidden Markov model method thatpredicts the presence of the SPs and their cleavage sites andalso discriminates such proteins from cytoplasmic and transmembraneones. The method performs satisfactorily, yielding a 35-foldcross-validation procedure, a sensitivity of 100% and specificity98.41% with the Matthews’ correlation coefficient beingequal to 0.964. This particular method is currently the onlyavailable method for the prediction of secretory SPs in archaea,and performs consistently and significantly better comparedwith other available predictors that were trained on sequencesof eukaryotic or bacterial origin. Searching 48 completely sequencedarchaeal genomes we identified 9437 putative SPs. The method,PRED-SIGNAL, and the results are freely available for academicusers at http://bioinformatics.biol.uoa.gr/PRED-SIGNAL/ andwe anticipate that it will be a valuable tool for the computationalanalysis of archaeal genomes.     

Analysis of multipath routing, part 2: mitigation of the effects of frequently changing network topologies

In this paper, we extend the analysis of multipath routing presented in our previous work, so that the basic restrictions on the evaluation and optimization of that scheme can be dropped (e.g., disjoint paths and identical paths in terms of failure probability). In that work, we employed diversity coding in order to provide increased protection against frequent route failures by splitting data packets and distributing them over multiple disjoint paths. Motivated by the high increase in the packet delivery ratio, we study the increase we can achieve through the usage of multiple paths in the general case, where the paths are not necessarily independent and their failure probabilities vary. For this reason, a function that measures the probability of successful transmission is derived as a tight approximation of the evaluation function P/sub succ/. Given the failure probabilities of the available paths and their correlation, we are able to find in polynomial time the set of paths that maximizes the probability of reconstructing the original information at the destination.