Hyperscale data centers are being deployed with tens of thousands of servers making operating efficiency a key metric. By optimizing the data center rack design, data center administrators can benefit from improved resource density, more flexible resource provisioning and reduced Total Cost of Ownership (TCO). Several industry initiatives have begun to drive standards for ...network infrastructure optimization. Facebook founded the Open Compute Project* several years ago to define standards around efficient server, storage and data center hardware for scalable computing. In China, Baidu, Alibaba, and Tencent joined forces and started Project Scorpio* in order to define efficient data center infrastructure standards. At Intel, we have been working for several years on a complimentary initiative called Intel® Rack Scale Architecture (Intel® RSA) which improves efficiency and deployment flexibility at the rack level while reducing TCO. This white paper will focus on the networking part of Intel RSA and show how our Intel® Ethernet Multi-host Controller FM10000 family can reduce cost and improve performance compared to standard rack solutions.