Font Size: a A A

A Research On Techniques Of Coded Caching In Linear Share Cache

Posted on:2022-08-04Degree:MasterType:Thesis
Country:ChinaCandidate:D Q LiangFull Text:PDF
GTID:2518306485486124Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Driven by the pursuit of high-quality multimedia contents by more and more wireless devices,the traffic of wireless network is increasing rapidly,and the wireless networks are increasingly under stress of data transmission.The high temporal variability of network traffic results in congestions during the peak traffic times and underutilization of the network during off-peak traffic times.By storing fractions of popular content at end users' local memories during the off-peak traffic times,coded caching creates the opportunity of coding multicast in the peak traffic times,thereby further reducing the transmission pressure of wireless network.In order to better study the coded caching,the concept of placement delivery array(PDA)is proposed,which is used to describe the coded caching scheme.The problem of studying coded caching scheme is transformed into the problem of designing a reasonable PDA.The network systems coded caching scheme formulated assume that each end user is equipped with a local cache.However in the actual scenario,the local memory capacity of the end user is quite small compared with the size of popular files.In addition,each end user can access network nodes nearby like macro base station and Content Delivery Network servers,such that it gets cached content which is much larger than its local memory that it can store.This paper will focus on a network scenario that allows end users to access multiple cache-nodes,named linear shared cache network.Each user can access multiple cache-nodes in the linear shared cache network.For the linear shared cache network,this paper proposes a transformation method,which is used to construct coded caching scheme under this network,resulting in a reasonable placement strategy and a delivery strategy with superior performance.The placement strategy makes full use of the cache capacity of cache-nodes,which is designed to ensure that the cache content of multiple consecutive cache-nodes will not be repeated.The designed delivery strategy can create multicast gain as much as possible,thus effectively reducing the transmission rate of the system.The proposed transformation method establishes the connection between PDAs and coded caching scheme for this network.Compared with the existing general coding caching schemes for this network scenario,the constructed coded caching scheme based on a class of PDA has the lowest transmission rate.Then,we prove that the proposed transformation method can also be applied to other types of PDA,and the transmission rate of the proposed scheme can be further reduced by using the coding idea.Decoding multicast signals needs to reveal the each user's request in the linear shared cache network,which would leaks the information about the private request of each user.Again,We formulate the linear shared cache network model under the privacy request problem.Then for this problem,we propose a novel method,which can transform the existing coded caching scheme satisfying some conditions into a coded caching scheme that can preserve the privacy of user's request.The key idea of the method is that treats the sum of user's request file and a linear combination as a virtual file,and then the virtual file is regarded as the user's request,which can hide the user's real request.Then from the perspective of information theory,we prove that the coded caching scheme constructed by the proposed method satisfies the privacy constraint.
Keywords/Search Tags:coded caching, placement delivery array, linear share cache, privacy, information theory
PDF Full Text Request
Related items