Cloud-Optimized Silicon Powers the Scalable Data Center | FMS 2022
Watch the Keynote presentation “Cloud-Optimized Silicon Powers the Scalable Data Center” with Marvell’s Jon Haswell, Senior Vice President, Firmware, and Gary Kotzur, CTO, Storage Products Group from the 2022 Flash Memory Summit. With the demand for workloads to scale up, down and everywhere in between, public cloud and hybrid cloud environments are under pressure to provide the ultimate in performance while at the same time remaining cost-competitive. Designing the highest performing storage and memory infrastructure for all workloads remains an expensive proposition at cloud scale, often resulting in underutilized resources – there has to be a better way. Optimizing NVMe and CXL-based silicon solutions from both a hardware and firmware perspective is proving the be the key to success in driving the foundation for the most scalable data center infrastructure.
Marvell’s Teralynx® Ethernet switch includes industry-unique telemetry features – high delay capture (HDC) and buffer drop capture (BDC) -- designed to provide high visibility into application performance and to feed details into open-source network operations tools such as Grafana. With HDC and BDC, data center operators get the application-flow data they need to highlight the flows experiencing delays or drops.
