Fatih Yazıcı, Ayhan Sefa Yıldız, Alper Yazar, E. G. Schmidt
{"title":"一种支持硬件加速云数据中心服务质量的新型可扩展片上交换机架构","authors":"Fatih Yazıcı, Ayhan Sefa Yıldız, Alper Yazar, E. G. Schmidt","doi":"10.1109/CloudNet51028.2020.9335788","DOIUrl":null,"url":null,"abstract":"This paper proposes a scalable on-chip packet switch architecture, ACCLOUD-SWITCH, for hardware accelerated cloud data centers. The proposed switch architecture adopts architectural features from high-speed computer network and network on chip (NoC) routers. ACCLOUD-SWITCH interconnects heterogeneous high-speed interfaces and is implemented on FPGA. The switch fabric runs at line speed for scalability. We propose a new work-conserving fabric arbiter that can allocate bandwidth to input/output pairs by prioritizing the switch ports and a new hybrid buffer structure for ports connected to reconfigurable regions for more efficient memory use. The switch is implemented for Xilinx Zynq SoC device to work at 40 Gbps. Our simulation results demonstrate the benefits of the proposed arbiter and the hybrid buffer structure.","PeriodicalId":156419,"journal":{"name":"2020 IEEE 9th International Conference on Cloud Networking (CloudNet)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2020-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":"{\"title\":\"A Novel Scalable On-chip Switch Architecture with Quality of Service Support for Hardware Accelerated Cloud Data Centers\",\"authors\":\"Fatih Yazıcı, Ayhan Sefa Yıldız, Alper Yazar, E. G. Schmidt\",\"doi\":\"10.1109/CloudNet51028.2020.9335788\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"This paper proposes a scalable on-chip packet switch architecture, ACCLOUD-SWITCH, for hardware accelerated cloud data centers. The proposed switch architecture adopts architectural features from high-speed computer network and network on chip (NoC) routers. ACCLOUD-SWITCH interconnects heterogeneous high-speed interfaces and is implemented on FPGA. The switch fabric runs at line speed for scalability. We propose a new work-conserving fabric arbiter that can allocate bandwidth to input/output pairs by prioritizing the switch ports and a new hybrid buffer structure for ports connected to reconfigurable regions for more efficient memory use. The switch is implemented for Xilinx Zynq SoC device to work at 40 Gbps. Our simulation results demonstrate the benefits of the proposed arbiter and the hybrid buffer structure.\",\"PeriodicalId\":156419,\"journal\":{\"name\":\"2020 IEEE 9th International Conference on Cloud Networking (CloudNet)\",\"volume\":\"1 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2020-11-09\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"2\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2020 IEEE 9th International Conference on Cloud Networking (CloudNet)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/CloudNet51028.2020.9335788\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2020 IEEE 9th International Conference on Cloud Networking (CloudNet)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/CloudNet51028.2020.9335788","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
A Novel Scalable On-chip Switch Architecture with Quality of Service Support for Hardware Accelerated Cloud Data Centers
This paper proposes a scalable on-chip packet switch architecture, ACCLOUD-SWITCH, for hardware accelerated cloud data centers. The proposed switch architecture adopts architectural features from high-speed computer network and network on chip (NoC) routers. ACCLOUD-SWITCH interconnects heterogeneous high-speed interfaces and is implemented on FPGA. The switch fabric runs at line speed for scalability. We propose a new work-conserving fabric arbiter that can allocate bandwidth to input/output pairs by prioritizing the switch ports and a new hybrid buffer structure for ports connected to reconfigurable regions for more efficient memory use. The switch is implemented for Xilinx Zynq SoC device to work at 40 Gbps. Our simulation results demonstrate the benefits of the proposed arbiter and the hybrid buffer structure.