OCP US Summit 2017 has ended

Sign up or log in to bookmark your favorites and sync them to your phone or calendar.

Engineering Workshop [clear filter]
Thursday, March 9


Welcome and Track Overview
avatar for John Stuewe

John Stuewe

ESI Architecture, Dell EMC
avatar for Siamak Tavallaei

Siamak Tavallaei

Principal Architect, Microsoft
Siamak Tavallaei is a Principal Architect at Microsoft’s Azure division. Collaborating with industry partners, he drives a number of initiatives in research, design, and deployment of hardware for Microsoft’s cloud-scale services such as Azure, Machine Learning, Bing, Office 365... Read More →

Thursday March 9, 2017 8:45am - 8:55am
Grand Ballroom C


OCP Mezzanine NIC 2.0 and Beyond
Update the status to the community about the latest about Mezz NIC. We will cover specification refresh to Gen4 PCIe, and discuss the options for expansion of form factor to support larger thermal profile and use cases in OCP community.

Link to slides: http://files.opencompute.org/oc/public.php?service=files&t=f6ba53f7bf13df7070fd4e1d10b6b11c

avatar for Jia Ning

Jia Ning

Hardware Engineering Manager, Facebook

Thursday March 9, 2017 9:00am - 9:55am
Grand Ballroom C


Multi-Node 1S Server Platform Yosemite V2
Update the community with latest 1S server platform development progress.

Link to slides: http://files.opencompute.org/oc/public.php?service=files&t=e1794e598ec0f98b535a85214f373ce0 


Thursday March 9, 2017 10:00am - 10:25am
Grand Ballroom C


3U High Density Compute Accelerator System (Dr. Know)

Dr. Know, a new EIA 19” 3U high performance compute accelerator, allows connection from one to four servers. For the 3U four drawers architecture, there are 16 high performance compute accelerator in 4 removable drawers and are compatible with the latest compute acceleration architectures.  The system provides higher computing performance and reliability to complete a large number of computing tasks. In addition to efficiency and reliability, Dr. Know also shows application flexibility in computing power sharing and various rack supporting. Furthermore, its tool-less concept eases maintenance of GPU cards, fan modules and PSUs.

Link to slides: http://files.opencompute.org/oc/public.php?service=files&t=6551fc367e20df0d099ca15f569065b1 

avatar for Terry Hou

Terry Hou


Thursday March 9, 2017 10:30am - 10:55am
Grand Ballroom C


Facebook OCP 2S Server Tioga Pass
Will talk about Facebook next generation Intel based 2S server, architecture, features etc.

Link to slides: http://files.opencompute.org/oc/public.php?service=files&t=a64a697b35f755abae1615ad1e08c208 

avatar for Whitney Zhao

Whitney Zhao

Hardware Engineer, Facebook Inc.
Whitney is a seasoned hardware engineer leading AI/ML system design in Facebook. Whitney has led multiple hardware generations ranging from general purpose 2S system such as Tioga Pass to ML JBOG Big Basin systems, all of which have been contributed to OCP. She has been driving multiple... Read More →

Thursday March 9, 2017 11:00am - 11:25am
Grand Ballroom C


AVA - NVMe m.2 in Scale Out Storage
This talk will overview Ava, an NVME m.2 carrier fit for various server form factors, allowing maximum flexibility and cost efficiencies when deploying flash at scale. This talk will overview the various Ava configurations supported, how they enable flexibility and savings, the type of workloads that benefit from this solution, how Facebook designed w/ MSFT, plans to deploy Ava, and more.

Link to slides: http://files.opencompute.org/oc/public.php?service=files&t=a843f1420b75eaad6ea05c4e4a700855 

avatar for Dominic Cheng

Dominic Cheng

Hardware Engineer, Facebook

Thursday March 9, 2017 11:30am - 11:55am
Grand Ballroom C


SmartNIC Data Plane Acceleration and Reconfiguration
Data centers desire flexibility within their architecture to respond to changing workloads and host software defined networking (SDN) requirements. The drive to higher performance networks like 25G and beyond means that CPU alone is not sufficient to enable flexible server-based networking. On top of this, kernel bypass can lead to a number of maintainability problems as keeping track with changes in the Linux kernel networking stack in user space requires significant development effort. In this talk we outline a kernel and upstream centric approach to data plane acceleration using an upstream SmartNIC BPF JIT. This allows extended Berkeley Packet Filter (eBPF) bytecode to be transparently offloaded to the SmartNIC from either the Traffic Control (TC) or Express Data Path (XDP) hooks in the kernel and could be used for applications such as DoS protection, Load balancing and software switching e.g Open vSwitch (OVS). We then follow this by outlining a proposed OCP contribution related to an open approach for reconfiguration using directly compiled SmartNIC programs in situations where BPF bytecode alone is not sufficient to accommodate changing semantics in the network.

Link to slides: http://files.opencompute.org/oc/public.php?service=files&t=c734ec68257253208dc8b7777c0d24d5 

avatar for Nic Viljoen

Nic Viljoen

Associate Director of Engineering, Netronome
Nic works on upstream solutions using SmartNICs, focusing on the interaction between SmartNICs and the Linux kernel using the upstream BPF JIT for the Netronome NFP processor.

Thursday March 9, 2017 1:00pm - 1:25pm
Grand Ballroom C


Zaius & Barreleye G2: the 48v Google / Rackspace OpenPOWER Platform
Learn the latest about Google & Rackspace's OCP / OpenPOWER collaborative effort. Includes implementation insights on 48-volt servers, OpenCAPI connectors and card form factors, and PCIe Gen4 OCP mezzanine solutions. Also includes a new / improved OpenBMC implementation.

Link to slides: http://files.opencompute.org/oc/public.php?service=files&t=89a0989114bbc44615ba7c691f1d5a24 

avatar for Adi Gangidi

Adi Gangidi

Senior Systems Design Engineer, Rackspace
Adi Gangidi is a Senior Systems Design Engineer at Rackspace, leading the Barreleye G2 program, an OpenPOWER / Open Compute initiative. He helps Rackspace build and scale, the high-performance, cost-effective, Barreleye / Zaius Servers , featuring IBM POWER 8/9 processors. He is excited... Read More →

Thursday March 9, 2017 1:30pm - 1:55pm
Grand Ballroom C


Big Basin: Flexible GPU Expander
This talk is to describe the architecture for the Big Basin, GPU expander. The talk includes the PCIe connectivity from an external CPU system, its system management and how it is deployed in the datacentre.

Link to slides: http://files.opencompute.org/oc/public.php?service=files&t=b9636dc6110a1bf2cd91314491647a57 

avatar for Koh Yew Thoon

Koh Yew Thoon

Hardware Engineer, Facebook, Inc.
Experienced in hardware design for past 15 years. Focus on high power/high speed digital systems. Have been working on chassis and stackable networking switches and started to work on Machine Learning servers in 2016.

Thursday March 9, 2017 2:00pm - 2:25pm
Grand Ballroom C


OCP Toolchain Subgroup - Ohub

avatar for Jean-Marie Verdun

Jean-Marie Verdun

CEO, Horizon Computing Solutions Corp
President, CEO, passionate by technology and efficiency Jean-Marie has participated to the expansion of major IT technologies during the past 20 years including the Alpha chip, highly parallel scientific clusters and free software.

Thursday March 9, 2017 2:30pm - 2:55pm
Grand Ballroom C