Howie Xu is VP of Engineering at Big Switch Networks, the leading platform-independent Open SDN company following nine years at VMware. In this keynote Howie will discuss the SDN industry landscape, how it impacts Data Center Networking, and Open SDN/ Network Virtualization vs. proprietary solutions.

2012 marked a critical milestone for Data Center infrastructure as the Xeon E5-series of processors and the world's first and only integrated 10GBASE-T MAC/PHY for LOM were jointly launched. For the first time, CIOs and IT organizations had access to systems with cost, power, density and latency per Gigabit/s metrics that justified the migration to the new interconnect technology. The presentation covers these trends, the metrics and the fast-growing 10GBASE-T ecosystem.

New Massively Scalable Data Centers (MSDCs) today host public and private clouds for companies like Amazon (EC2), Apple(iCloud) and Verizon Terremarkoffering services like SaaS, PaaS and IaaS. At 10 times the size of their predecessors, these data centers require a whole new network paradigm, needing exponentially more bandwidth while pushing the limits of power and cost per bit. This presentation explores how 100G CMOS connectivity chips will play a critical role in the deployment of 100G networks in these data centers.

A new task force has been formed to develop the latest 100Gstandard - 802.3bm. This presentation will provide insight into thecurrent direction of the IEEE 802.3bm standard activity, including technicalhighlights on the various proposals under consideration.

SDN forwarding engines or complex security applications require memory access rates to support 600Mpps throughput. On-chip memory lacks capacity to hold increasing tables, rule sets, meters and statistics counters. Second generation Bandwidth Engine architecture supports high throughput, low latency firewalls, monitoring switches, SDNs, or traditional core networks. In this discussion, MoSys introduces the newest member of the Bandwidth Engine 2 family, which delivers over 4GA/s, up to 192Gbps duplex throughput and intelligent offload capabilities while reducing system-level power and cost.

With DDR3/4 interface and DRAM plus NAND components, Nonvolatile dual in line memory modules (NVDIMMs) are enabling the growth in enterprise and cloud computing.NVDIMMs enables data recovery in the event of a power failure; system check pointing; and even SSDs with a DDR3/4 interface. This presentation will show how NVDIMMs bridge a gap between today's DRAM technology and tomorrow's nonvolatile DRAM replacement

There has been a lot of buzz about the potential energy savings by ARM processors in servers, but there has been little data. This session will feature analyses of production-ready ARM server performance under various workloads, showing where ARM servers can meet the needs of the market, and where they still need some work. The talk will conclude with a view of the future of ARM, with details about the Calxeda Energycore roadmap.

As the data center compute infrastructure is shifting to meet the demands of evolving workloads, a new server architecture has emerged. This presentation will describe how a new category of cloud server, enabled by the ARM 64-bit architecture and the X-Gene Server on a Chip, will deliver the performance, power efficiency, and integration needed to address evolving workloads in the data center.

This presentation will provide an overview of the server processors available on the market and the key attributes of different architectures.