{"title":"VBR业务的VLSI ATM交换机架构","authors":"N. Ranganathan, R. Anand, G. Chiruvolu","doi":"10.1109/ICVD.1998.646644","DOIUrl":null,"url":null,"abstract":"An Asynchronous Transfer Mode (ATM) switching network must process data at the rates of 155 Mbps and 620 Mbps as per the standard. Such bandwidth requirements have necessitated the realization of efficient switch architectures. In this paper, we propose a novel architecture for the design of a non-blocking, central buffer switch for ATM networks. The central buffer switch architectures described in the literature organize the logical output queues as linked lists of data packets. Thus, dynamic memory allocation involves the manipulation of the read and the write pointers of these linked lists. In the switch architecture proposed in this work, the packets are stored in the data memory and only the packet addresses are stored in a set of First In First Out (FIFO) buffers that form the logical output queues. This approach eliminates the need for memory accesses for the manipulation of linked lists which improves significantly the response time. A 4/spl times/4 prototype switch of the proposed architecture was designed and verified using the Cadence design tools. The prototype was verified to operate at a frequency of 40 MHz yielding a throughput of 12.334 Gbps.","PeriodicalId":139023,"journal":{"name":"Proceedings Eleventh International Conference on VLSI Design","volume":"62 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1998-01-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"6","resultStr":"{\"title\":\"A VLSI ATM switch architecture for VBR traffic\",\"authors\":\"N. Ranganathan, R. Anand, G. Chiruvolu\",\"doi\":\"10.1109/ICVD.1998.646644\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"An Asynchronous Transfer Mode (ATM) switching network must process data at the rates of 155 Mbps and 620 Mbps as per the standard. Such bandwidth requirements have necessitated the realization of efficient switch architectures. In this paper, we propose a novel architecture for the design of a non-blocking, central buffer switch for ATM networks. The central buffer switch architectures described in the literature organize the logical output queues as linked lists of data packets. Thus, dynamic memory allocation involves the manipulation of the read and the write pointers of these linked lists. In the switch architecture proposed in this work, the packets are stored in the data memory and only the packet addresses are stored in a set of First In First Out (FIFO) buffers that form the logical output queues. This approach eliminates the need for memory accesses for the manipulation of linked lists which improves significantly the response time. A 4/spl times/4 prototype switch of the proposed architecture was designed and verified using the Cadence design tools. The prototype was verified to operate at a frequency of 40 MHz yielding a throughput of 12.334 Gbps.\",\"PeriodicalId\":139023,\"journal\":{\"name\":\"Proceedings Eleventh International Conference on VLSI Design\",\"volume\":\"62 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"1998-01-04\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"6\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings Eleventh International Conference on VLSI Design\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICVD.1998.646644\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings Eleventh International Conference on VLSI Design","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICVD.1998.646644","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 6
摘要
ATM (Asynchronous Transfer Mode)交换网络的数据处理速率必须达到155mbps和620mbps的标准。这样的带宽需求使得实现高效的交换机架构成为必要。在本文中,我们提出了一种用于ATM网络的非阻塞中央缓冲交换机的新架构。文献中描述的中央缓冲交换机体系结构将逻辑输出队列组织为数据包的链表。因此,动态内存分配涉及到对这些链表的读和写指针的操作。在这项工作中提出的交换机架构中,数据包存储在数据存储器中,只有数据包地址存储在形成逻辑输出队列的一组先进先出(FIFO)缓冲区中。这种方法消除了为操作链表而访问内存的需要,从而显著提高了响应时间。使用Cadence设计工具设计并验证了所提出架构的4/ sp1倍/4原型开关。经过验证,该原型在40 MHz的频率下运行,吞吐量为12.334 Gbps。
An Asynchronous Transfer Mode (ATM) switching network must process data at the rates of 155 Mbps and 620 Mbps as per the standard. Such bandwidth requirements have necessitated the realization of efficient switch architectures. In this paper, we propose a novel architecture for the design of a non-blocking, central buffer switch for ATM networks. The central buffer switch architectures described in the literature organize the logical output queues as linked lists of data packets. Thus, dynamic memory allocation involves the manipulation of the read and the write pointers of these linked lists. In the switch architecture proposed in this work, the packets are stored in the data memory and only the packet addresses are stored in a set of First In First Out (FIFO) buffers that form the logical output queues. This approach eliminates the need for memory accesses for the manipulation of linked lists which improves significantly the response time. A 4/spl times/4 prototype switch of the proposed architecture was designed and verified using the Cadence design tools. The prototype was verified to operate at a frequency of 40 MHz yielding a throughput of 12.334 Gbps.