Pub Date : 2015-09-01DOI: 10.1109/ALLERTON.2015.7447025
Javad Ghaderi
We consider the problem of scheduling VMs (Virtual Machines) in a multi-server system motivated by cloud computing applications. VMs arrive dynamically over time and require various amounts of resources (e.g., CPU, Memory, Storage, etc.) for the duration of their service. When a VM arrives, it is queued and later served by one of the servers that has sufficient remaining capacity to serve it. The scheduling of VMs is subject to: (i) packing constraints, i.e., multiple VMs can be served simultaneously by a single server if their cumulative resource requirement does not violate the capacity of the server, and (ii) non-preemption, i.e., once a VM is scheduled in a server, it cannot be interrupted or migrated to another server. To achieve maximum throughput, prior results hinge on solving a hard combinatorial problem (Knapsack) at the instances that all the servers become empty (the so-called global refresh times which require synchronization among the servers). The main contribution of this paper is that it resolves these issues. Specifically, we present a class of randomized algorithms for placing VMs in the servers that can achieve maximum throughput without preemptions. The algorithms are naturally distributed, have low complexity, and each queue needs to perform limited operations. Further, our algorithms display good delay performance in simulations, comparable to delay of heuristics that may not be throughput-optimal, and much better than the delay of the prior known throughput-optimal algorithms.
{"title":"Simple high-performance algorithms for scheduling jobs in the cloud","authors":"Javad Ghaderi","doi":"10.1109/ALLERTON.2015.7447025","DOIUrl":"https://doi.org/10.1109/ALLERTON.2015.7447025","url":null,"abstract":"We consider the problem of scheduling VMs (Virtual Machines) in a multi-server system motivated by cloud computing applications. VMs arrive dynamically over time and require various amounts of resources (e.g., CPU, Memory, Storage, etc.) for the duration of their service. When a VM arrives, it is queued and later served by one of the servers that has sufficient remaining capacity to serve it. The scheduling of VMs is subject to: (i) packing constraints, i.e., multiple VMs can be served simultaneously by a single server if their cumulative resource requirement does not violate the capacity of the server, and (ii) non-preemption, i.e., once a VM is scheduled in a server, it cannot be interrupted or migrated to another server. To achieve maximum throughput, prior results hinge on solving a hard combinatorial problem (Knapsack) at the instances that all the servers become empty (the so-called global refresh times which require synchronization among the servers). The main contribution of this paper is that it resolves these issues. Specifically, we present a class of randomized algorithms for placing VMs in the servers that can achieve maximum throughput without preemptions. The algorithms are naturally distributed, have low complexity, and each queue needs to perform limited operations. Further, our algorithms display good delay performance in simulations, comparable to delay of heuristics that may not be throughput-optimal, and much better than the delay of the prior known throughput-optimal algorithms.","PeriodicalId":112948,"journal":{"name":"2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116418832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-09-01DOI: 10.1109/ALLERTON.2015.7447056
Andrey Gushchin, A. Elwalid, A. Tang
Routing in SDN-enabled networks with middle-boxes enables high level of control over network flows so that desired level of security, efficiency and performance can be achieved. However, the new capabilities place constraints on available resources such as switch memory and middle-box processing capacity. Additionally, flow traversal constraints may also arise where each network flow has to visit certain middleboxes in a specific order. In this work, we investigate the problem of achieving global routing objectives while satisfying necessary constraints. We formulate the routing problem as an integer linear optimization problem, and propose an efficient randomized algorithm for solving it. We provide both theoretical and practical (using simulations) analysis for our algorithm. The simulations show that while other routing approaches (including greedy and previously proposed one) violate the switch memory capacity constraints (requiring 10 times more capacity in some cases), and could attain a low objective value, our algorithm satisfies all traversal and network resource constraints and achieves a high objective value.
{"title":"Enabling service function chaining through routing optimization in software defined networks","authors":"Andrey Gushchin, A. Elwalid, A. Tang","doi":"10.1109/ALLERTON.2015.7447056","DOIUrl":"https://doi.org/10.1109/ALLERTON.2015.7447056","url":null,"abstract":"Routing in SDN-enabled networks with middle-boxes enables high level of control over network flows so that desired level of security, efficiency and performance can be achieved. However, the new capabilities place constraints on available resources such as switch memory and middle-box processing capacity. Additionally, flow traversal constraints may also arise where each network flow has to visit certain middleboxes in a specific order. In this work, we investigate the problem of achieving global routing objectives while satisfying necessary constraints. We formulate the routing problem as an integer linear optimization problem, and propose an efficient randomized algorithm for solving it. We provide both theoretical and practical (using simulations) analysis for our algorithm. The simulations show that while other routing approaches (including greedy and previously proposed one) violate the switch memory capacity constraints (requiring 10 times more capacity in some cases), and could attain a low objective value, our algorithm satisfies all traversal and network resource constraints and achieves a high objective value.","PeriodicalId":112948,"journal":{"name":"2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117088279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-09-01DOI: 10.1109/ALLERTON.2015.7447180
C. Pehlevan, D. Chklovskii
In analyzing information streamed by sensory organs, our brains face challenges similar to those solved in statistical signal processing. This suggests that biologically plausible implementations of online signal processing algorithms may model neural computation. Here, we focus on such workhorses of signal processing as Principal Component Analysis (PCA) and whitening which maximize information transmission in the presence of noise. We adopt the similarity matching framework, recently developed for principal subspace extraction, but modify the existing objective functions by adding a decorrelating term. From the modified objective functions, we derive online PCA and whitening algorithms which are implementable by neural networks with local learning rules, i.e. synaptic weight updates that depend on the activity of only pre- and postsynaptic neurons. Our theory offers a principled model of neural computations and makes testable predictions such as the dropout of underutilized neurons.
{"title":"Optimization theory of Hebbian/anti-Hebbian networks for PCA and whitening","authors":"C. Pehlevan, D. Chklovskii","doi":"10.1109/ALLERTON.2015.7447180","DOIUrl":"https://doi.org/10.1109/ALLERTON.2015.7447180","url":null,"abstract":"In analyzing information streamed by sensory organs, our brains face challenges similar to those solved in statistical signal processing. This suggests that biologically plausible implementations of online signal processing algorithms may model neural computation. Here, we focus on such workhorses of signal processing as Principal Component Analysis (PCA) and whitening which maximize information transmission in the presence of noise. We adopt the similarity matching framework, recently developed for principal subspace extraction, but modify the existing objective functions by adding a decorrelating term. From the modified objective functions, we derive online PCA and whitening algorithms which are implementable by neural networks with local learning rules, i.e. synaptic weight updates that depend on the activity of only pre- and postsynaptic neurons. Our theory offers a principled model of neural computations and makes testable predictions such as the dropout of underutilized neurons.","PeriodicalId":112948,"journal":{"name":"2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)","volume":"24 6","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120868722","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-09-01DOI: 10.1109/ALLERTON.2015.7447002
S. Manaffam, A. Seyedi, A. Vosoughi, T. Javidi
Here, we study the ultimately bounded stability of network of mismatched systems using Lyapunov direct method. The upper bound on the error of oscillators from the center of the neighborhood is derived. Then the performance of an adaptive compensation via decentralized control is analyzed. Finally, the analytical results for a network of globally connected Lorenz oscillators are verified.
{"title":"Bounded stability in networked systems with parameter mismatch and adaptive decentralized estimation","authors":"S. Manaffam, A. Seyedi, A. Vosoughi, T. Javidi","doi":"10.1109/ALLERTON.2015.7447002","DOIUrl":"https://doi.org/10.1109/ALLERTON.2015.7447002","url":null,"abstract":"Here, we study the ultimately bounded stability of network of mismatched systems using Lyapunov direct method. The upper bound on the error of oscillators from the center of the neighborhood is derived. Then the performance of an adaptive compensation via decentralized control is analyzed. Finally, the analytical results for a network of globally connected Lorenz oscillators are verified.","PeriodicalId":112948,"journal":{"name":"2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124480870","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-09-01DOI: 10.1109/ALLERTON.2015.7446995
Haewon Jeong, P. Grover
We provide the first constructions of a new family of error-correcting codes called “energy-adaptive codes.” These codes are designed to enable adaptive circuit implementations that minimize the total system-level energy based on varying distances and target error probabilities. Recent work has explored fundamental limits and practical strategies for minimizing total (transmit + circuit) power, considering both power consumed in computational nodes as well as wires in the circuit. It is now established that to minimize total power, code choice and circuit-design should change with communication distances and/or target error probability. Motivated by circuit area constraints, the energy-adaptive codes adapt energy consumption as distances and/or target error probability change. These codes shrink and expand the wiring area they occupy as demands on the system change, adjusting the hardware by turning on and off non-local wires in the circuit. The code constructions are hierarchical, and use QC-LDPC codes as the basis. We estimate the decoding power savings attained by use of these codes through simulation results. Such codes can be of increasing utility as we enter the era of 1000 devices per person where designing the skilled labor for obtaining optimized designs for each system will simply be unavailable. While our first constructions are admittedly simplistic, the goal of the paper is to bring this new problem to the attention of the coding-theory community.
{"title":"Energy-adaptive codes","authors":"Haewon Jeong, P. Grover","doi":"10.1109/ALLERTON.2015.7446995","DOIUrl":"https://doi.org/10.1109/ALLERTON.2015.7446995","url":null,"abstract":"We provide the first constructions of a new family of error-correcting codes called “energy-adaptive codes.” These codes are designed to enable adaptive circuit implementations that minimize the total system-level energy based on varying distances and target error probabilities. Recent work has explored fundamental limits and practical strategies for minimizing total (transmit + circuit) power, considering both power consumed in computational nodes as well as wires in the circuit. It is now established that to minimize total power, code choice and circuit-design should change with communication distances and/or target error probability. Motivated by circuit area constraints, the energy-adaptive codes adapt energy consumption as distances and/or target error probability change. These codes shrink and expand the wiring area they occupy as demands on the system change, adjusting the hardware by turning on and off non-local wires in the circuit. The code constructions are hierarchical, and use QC-LDPC codes as the basis. We estimate the decoding power savings attained by use of these codes through simulation results. Such codes can be of increasing utility as we enter the era of 1000 devices per person where designing the skilled labor for obtaining optimized designs for each system will simply be unavailable. While our first constructions are admittedly simplistic, the goal of the paper is to bring this new problem to the attention of the coding-theory community.","PeriodicalId":112948,"journal":{"name":"2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125794696","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-09-01DOI: 10.1109/ALLERTON.2015.7447099
Jun Zhao
One-dimensional geometric random graphs are constructed by distributing n nodes uniformly and independently on a unit interval and then assigning an undirected edge between any two nodes that have a distance at most τn. These graphs have received much interest and been used in various applications including wireless networks. A threshold of τn for connectivity is known as τ*n = ln n/n in the literature. In this paper, we prove that a threshold of τn for the absence of isolated node is ln n/2n (i.e., a half of the threshold τ*n). Our result shows there is a gap between thresholds of connectivity and the absence of isolated node in one-dimensional geometric random graphs; in particular, when τn equals c ln n/n for a constant c ∈ (1/2, 1), a one-dimensional geometric random graph has no isolated node but is not connected. This gap in one-dimensional geometric random graphs is in sharp contrast to the prevalent phenomenon in many other random graphs such as two-dimensional geometric random graphs, Erdös-Rényi graphs, and random intersection graphs, all of which in the asymptotic sense become connected as soon as there is no isolated node.
一维几何随机图是通过在单位间隔上均匀独立地分布n个节点，然后在任意两个距离不超过τn的节点之间分配一条无向边来构造的。这些图形引起了人们的极大兴趣，并被用于包括无线网络在内的各种应用中。在文献中，连通性的阈值τn被称为τ*n = lnn /n。在本文中，我们证明了不存在孤立节点的阈值τn是ln n/2n(即阈值τ*n的一半)。研究结果表明:一维几何随机图的连通阈值与孤立节点缺失阈值之间存在差距;特别地，对于常数c∈(1/ 2,1)，当τn = c ln n/n时，一维几何随机图没有孤立节点，但不连通。一维几何随机图中的这种差距与许多其他随机图中的普遍现象形成鲜明对比，例如二维几何随机图、Erdös-Rényi图和随机交点图，所有这些图在渐近意义上只要没有孤立节点就会相互连接。
{"title":"The absence of isolated node in geometric random graphs","authors":"Jun Zhao","doi":"10.1109/ALLERTON.2015.7447099","DOIUrl":"https://doi.org/10.1109/ALLERTON.2015.7447099","url":null,"abstract":"One-dimensional geometric random graphs are constructed by distributing n nodes uniformly and independently on a unit interval and then assigning an undirected edge between any two nodes that have a distance at most τ<sub>n</sub>. These graphs have received much interest and been used in various applications including wireless networks. A threshold of τ<sub>n</sub> for connectivity is known as τ*<sub>n</sub> = ln n/n in the literature. In this paper, we prove that a threshold of τ<sub>n</sub> for the absence of isolated node is ln n/2n (i.e., a half of the threshold τ*<sub>n</sub>). Our result shows there is a gap between thresholds of connectivity and the absence of isolated node in one-dimensional geometric random graphs; in particular, when τ<sub>n</sub> equals c ln n/n for a constant c ∈ (1/2, 1), a one-dimensional geometric random graph has no isolated node but is not connected. This gap in one-dimensional geometric random graphs is in sharp contrast to the prevalent phenomenon in many other random graphs such as two-dimensional geometric random graphs, Erdös-Rényi graphs, and random intersection graphs, all of which in the asymptotic sense become connected as soon as there is no isolated node.","PeriodicalId":112948,"journal":{"name":"2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129471330","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-09-01DOI: 10.1109/ALLERTON.2015.7447125
Wenbo He, B. Nazer, S. Shamai
Consider a Gaussian multiple-input multiple-output (MIMO) multiple-access channel (MAC) with channel matrix H and a Gaussian MIMO broadcast channel (BC) with channel matrix Hτ. For the MIMO MAC, the integer-forcing architecture consists of first decoding integer-linear combinations of the transmitted codewords, which are then solved for the original messages. For the MIMO BC, the integer-forcing architecture consists of pre-inverting the integer-linear combinations at the transmitter so that each receiver can obtain its desired codeword by decoding an integer-linear combination. In both cases, integer-forcing offers higher achievable rates than zero-forcing. In recent work, we established an uplink-downlink duality relationship for integer-forcing, i.e., we showed that any rate tuple that is achievable via integer-forcing on the MIMO MAC can be achieved via integer-forcing on the MIMO BC with the same sum power and vice versa. It has also been shown that integer-forcing for the MIMO MAC can be enhanced via successive cancellation. Here, we introduce dirty-paper integer-forcing for the MIMO BC and establish uplink-downlink duality with successive integer-forcing for the MIMO MAC.
{"title":"Dirty-paper integer-forcing","authors":"Wenbo He, B. Nazer, S. Shamai","doi":"10.1109/ALLERTON.2015.7447125","DOIUrl":"https://doi.org/10.1109/ALLERTON.2015.7447125","url":null,"abstract":"Consider a Gaussian multiple-input multiple-output (MIMO) multiple-access channel (MAC) with channel matrix H and a Gaussian MIMO broadcast channel (BC) with channel matrix Hτ. For the MIMO MAC, the integer-forcing architecture consists of first decoding integer-linear combinations of the transmitted codewords, which are then solved for the original messages. For the MIMO BC, the integer-forcing architecture consists of pre-inverting the integer-linear combinations at the transmitter so that each receiver can obtain its desired codeword by decoding an integer-linear combination. In both cases, integer-forcing offers higher achievable rates than zero-forcing. In recent work, we established an uplink-downlink duality relationship for integer-forcing, i.e., we showed that any rate tuple that is achievable via integer-forcing on the MIMO MAC can be achieved via integer-forcing on the MIMO BC with the same sum power and vice versa. It has also been shown that integer-forcing for the MIMO MAC can be enhanced via successive cancellation. Here, we introduce dirty-paper integer-forcing for the MIMO BC and establish uplink-downlink duality with successive integer-forcing for the MIMO MAC.","PeriodicalId":112948,"journal":{"name":"2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129756391","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-09-01DOI: 10.1109/ALLERTON.2015.7447011
S. Sundaram, B. Gharesifard
We investigate the vulnerabilities of consensus-based distributed optimization protocols to nodes that deviate from the prescribed update rule (e.g., due to failures or adversarial attacks). After characterizing certain fundamental limitations on the performance of any distributed optimization algorithm in the presence of adversaries, we propose a robust consensus-based distributed optimization algorithm that is guaranteed to converge to the convex hull of the set of minimizers of the non-adversarial nodes' functions. We also study the distance-to-optimality properties of our proposed robust algorithm in terms of F-local sets of the graph. We show that finding the largest size of such sets is NP-hard.
{"title":"Consensus-based distributed optimization with malicious nodes","authors":"S. Sundaram, B. Gharesifard","doi":"10.1109/ALLERTON.2015.7447011","DOIUrl":"https://doi.org/10.1109/ALLERTON.2015.7447011","url":null,"abstract":"We investigate the vulnerabilities of consensus-based distributed optimization protocols to nodes that deviate from the prescribed update rule (e.g., due to failures or adversarial attacks). After characterizing certain fundamental limitations on the performance of any distributed optimization algorithm in the presence of adversaries, we propose a robust consensus-based distributed optimization algorithm that is guaranteed to converge to the convex hull of the set of minimizers of the non-adversarial nodes' functions. We also study the distance-to-optimality properties of our proposed robust algorithm in terms of F-local sets of the graph. We show that finding the largest size of such sets is NP-hard.","PeriodicalId":112948,"journal":{"name":"2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129050729","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-09-01DOI: 10.1109/ALLERTON.2015.7447178
Asma Ghorbel, M. Kobayashi, Sheng Yang
We consider a cache-enabled K-user broadcast erasure packet channel in which a server with a library of N files wishes to deliver a requested file to each user who is equipped with a cache of a finite memory M. Assuming that the transmitter has state feedback and user caches can be filled during off-peak hours reliably by decentralized cache placement, we characterize the optimal rate region as a function of the memory size, the erasure probability. The proposed delivery scheme, based on the scheme proposed by Gatzianas et al., exploits the receiver side information established during the placement phase. Our results enable us to quantify the net benefits of decentralized coded caching in the presence of erasure. The role of state feedback is found useful especially when the erasure probability is large and/or the normalized memory size is small.
{"title":"Cache-enabled broadcast packet erasure channels with state feedback","authors":"Asma Ghorbel, M. Kobayashi, Sheng Yang","doi":"10.1109/ALLERTON.2015.7447178","DOIUrl":"https://doi.org/10.1109/ALLERTON.2015.7447178","url":null,"abstract":"We consider a cache-enabled K-user broadcast erasure packet channel in which a server with a library of N files wishes to deliver a requested file to each user who is equipped with a cache of a finite memory M. Assuming that the transmitter has state feedback and user caches can be filled during off-peak hours reliably by decentralized cache placement, we characterize the optimal rate region as a function of the memory size, the erasure probability. The proposed delivery scheme, based on the scheme proposed by Gatzianas et al., exploits the receiver side information established during the placement phase. Our results enable us to quantify the net benefits of decentralized coded caching in the presence of erasure. The role of state feedback is found useful especially when the erasure probability is large and/or the normalized memory size is small.","PeriodicalId":112948,"journal":{"name":"2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131307080","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-09-01DOI: 10.1109/ALLERTON.2015.7447048
Youlong Wu
Coding scheme for discrete memoryless multicast networks with rate-limited feedback from the receivers and relays to the transmitter is proposed. The coding scheme is based on block-Markov coding, joint backward decoding and hybrid relaying strategy. In each block, the relays use partial decode-forward strategy to decode part of the source message. Meanwhile, the receivers and relays use compress-forward strategy to compress their channel outputs and send the compression indices to the transmitter through the feedback links. In the next block, after obtaining the compression indices, the transmitter sends them together with the source message. Each receiver uses backward decoding to jointly decode the source message and all compression indices. It is shown that our coding scheme generalizes Gabbai and Bross's results for the single relay channel with partial feedback, where they proposed coding schemes based on restricted decoding and deterministic partitioning. For the single relay channel with relay-transmitter feedback, our coding scheme can strictly improve on noisy network coding, distributed decode-forward coding and all known lower bounds on the achievable rate in the absence of feedback. Furthermore, motivated by the feedback coding scheme, we propose a new coding scheme for discrete memoryless multicast networks without feedback, which also improves noisy network coding and distributed decode-forward coding.
{"title":"Coding schemes for discrete memoryless multicast networks with and without feedback","authors":"Youlong Wu","doi":"10.1109/ALLERTON.2015.7447048","DOIUrl":"https://doi.org/10.1109/ALLERTON.2015.7447048","url":null,"abstract":"Coding scheme for discrete memoryless multicast networks with rate-limited feedback from the receivers and relays to the transmitter is proposed. The coding scheme is based on block-Markov coding, joint backward decoding and hybrid relaying strategy. In each block, the relays use partial decode-forward strategy to decode part of the source message. Meanwhile, the receivers and relays use compress-forward strategy to compress their channel outputs and send the compression indices to the transmitter through the feedback links. In the next block, after obtaining the compression indices, the transmitter sends them together with the source message. Each receiver uses backward decoding to jointly decode the source message and all compression indices. It is shown that our coding scheme generalizes Gabbai and Bross's results for the single relay channel with partial feedback, where they proposed coding schemes based on restricted decoding and deterministic partitioning. For the single relay channel with relay-transmitter feedback, our coding scheme can strictly improve on noisy network coding, distributed decode-forward coding and all known lower bounds on the achievable rate in the absence of feedback. Furthermore, motivated by the feedback coding scheme, we propose a new coding scheme for discrete memoryless multicast networks without feedback, which also improves noisy network coding and distributed decode-forward coding.","PeriodicalId":112948,"journal":{"name":"2015 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton)","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125504765","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：