{"title":"Reducing router's area in NoC by changing buffering method while providing QoS","authors":"M. Shariat, Nima Azizibabani","doi":"10.1109/WISES.2010.5548437","DOIUrl":null,"url":null,"abstract":"One of the most significant concerns in designing on-chip networks is guaranteeing QoS requirements for various applications. In this paper, we propose a buffering method which reduces the router's area and in turn network cost without affecting the QoS issues. The proposed method rather than considering a dedicated buffer for each class of traffic, assigns a single shared buffer on each port to the all existing prioritized traffic classes. Simulation results show that exploiting shared buffer increases buffer utilization. Moreover, the required space of the shared buffer is 37.5% less than the space needed for the dedicated buffers. Furthermore, a controller is developed to both manage the shared buffer and allocate a particular amount of the shared buffer as an upper limit to each class of traffic; accordingly, QoS requirements will be preserved for high prioritized traffic classes.","PeriodicalId":166416,"journal":{"name":"2010 8th Workshop on Intelligent Solutions in Embedded Systems","volume":"82 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2010-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2010 8th Workshop on Intelligent Solutions in Embedded Systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/WISES.2010.5548437","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
One of the most significant concerns in designing on-chip networks is guaranteeing QoS requirements for various applications. In this paper, we propose a buffering method which reduces the router's area and in turn network cost without affecting the QoS issues. The proposed method rather than considering a dedicated buffer for each class of traffic, assigns a single shared buffer on each port to the all existing prioritized traffic classes. Simulation results show that exploiting shared buffer increases buffer utilization. Moreover, the required space of the shared buffer is 37.5% less than the space needed for the dedicated buffers. Furthermore, a controller is developed to both manage the shared buffer and allocate a particular amount of the shared buffer as an upper limit to each class of traffic; accordingly, QoS requirements will be preserved for high prioritized traffic classes.