روش هاي كشينگ مبتني بر كدينگ منبع

چكيده انگليسي :

Source Coding Based Caching Methods Niloufar Ahmadypour n ahmadypour@ec iut ac ir Date of Submission 19 9 2015 Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan 84156 83111 Iran Degree M Sc Language PersianSupervisor Dr Mohammadali Khosravifard khosravi@cc iut ac ir Department of Electrical and Computer Engineering Isfahan University of TechnologyAbstract Congestion of users demands during peak traffic times forces the link to transfer a large amount of data with a hightransmission rate in a short time and to be useless for a long time This problem causes inefficiency of using thenetwork One method for overcoming this problem and reducing the transmission rate during the peak hours is Caching In this method when the number of the requests is small some contents are stored in memories caches which have been distributed in the network Placement phase and during peak hours will provide part of the users requests In the second phase Delivery phase the server receives the demands and sends out a common message toall users so that they can recover their requested content from this message The bottle neck in the placement phaseis the size of the cache memories in the network and in the delivery phase is the rate required to serve therequested content Till now caching has been studied from different viewpoints Some studies have been focusedon the delivery phase and the others have focused on the placement phase With the first approach for fixed cachecontents and specified demands the delivery phase is optimized These studies are almost similar to the sourcecoding problem with side information In fact the message must be coded such that all users can recover theirdemands by using the coded message and the available information in their own caches as the side information Index coding is the most important achievement with this approach In the index coding problem there are somereceivers that have some bits of the other receivers in their cache and try to recover their desired message from thecode which is sent by the server The side information graph indicates the bits which are stored in the cache of eachreceiver In another approach optimization of both phases is considered These studies are focused on the use ofcaching in specific applications such as video on demand VoD In this thesis in addition to reviewing the codedcaching and decentralized coded caching methods as the new achievements in this area the main focus is on thestudy of index coding In particular the importance of linear index coding and its relationship with network codingproblem is reviewed The main question that is raised in this study and the author seeks to answer is if some delay istolerable such that users can wait for recovering two or more consecutive requests together instead of sequentialrecovering can code length be reduced by index coding of two or more side information graphs in a jointlymanner Our observations shows that in some cases a significant code length reduction approximately 50 ispossible Specifically an upper bound on the length of joint index code is derived in the case that union of sideinformation graphs is the complete graph On the other hand there are conditions in terms of side informationgraphs that joint index coding does not result in any advantage in comparison with the conventional index coding Keywords Coded Caching Source Coding with Side Information Index Coding Joint Index Coding

احمدي پور، نيلوفر

روش هاي كشينگ مبتني بر كدينگ منبع

كدينگ منبع با اطلاعات جانبي , ايندكس - كدينگ , ايندكس - كدينگ توام