Tuesday, August 9, 2016

Open vSwitch (OVS) Moves to Linux Foundation

Open vSwitch (OVS), which is an open source virtual switch designed to enable network automation while supporting standard management interfaces and protocols, is now a Linux Foundation Project.

OVS was developed to address the rapidly growing needs of SDN and virtual networking use cases. It has been ported to multiple virtualization platforms, switching chipsets, and networking hardware accelerators. OVS works on a wide variety of systems, including Linux, DPDK, Hyper-V, and FreeBSD. It is used in a variety of SDN applications, including NFV and network virtualization; it is the most widely used networking back-end in OpenStack.

“OVS is a great example of how open source software has enabled the networking industry to match the pace of cloud computing and help advance virtualized technologies,” said Jim Zemlin, executive director, The Linux Foundation. “Hosting OVS as a Linux Foundation Project will serve to further collaboration across users and vendors and aid in open technology development throughout the networking stack.”

Contributions to OVS span more than 300 individuals across companies including Cisco, Ericsson, Huawei, HP, IBM, Intel, Red Hat and VMware. The governance is managed by a group of the top committers to the project with a variety of backgrounds and affiliations.

As a project hosted at The Linux Foundation, Open vSwitch will continue its operations under an open governance model and continue accepting contributions from all interested companies and developers.



Momentum Builds for Central Office Re-architected as a Data Center Project (CORD)

The Central Office Re-architected as a Data Center (CORD) initiative, which was initially developed by ON.Lab as a use case for the ONOS open source SDN operating system, will now be managed as an independent open source project under The Linux Foundation. CORD aims to utilize merchant silicon and  the elasticity of commodity clouds to enable data center economics and cloud agility in the central office environment. Google, Radisys and Samsung...

Facebook Pursues Hierarchical Flash Storage Strategy

Building data center infrastructure at Facebook is all about understanding engagement, said Vijay Rao, Director Technology, Strategy Facebook, speaking at Flash Memory Summit in Santa Clara, California.

Facebook's growth metrics are well known - whether in posts, comments, likes, photos, messages, videos -- the curve moves strongly in the upward and to the right direction.

Designing infrastructure for this type of runaway engagement means planning for spikes, such as Mother's Day.  In terms of raw power, Rao estimated Facebook's overall compute power at about 7.5 quadrillion instructions per second. Facebook engineers now have seven types of servers to choose from depending on the application they are supporting.

On the storage side, Facebook is pursuing a hierarchical storage model -- various technologies are employed depending on the "temperature" of the data.  Hot data resides in DRAM or NVM-based DIMMs while warm data is stored in PCIe NVM Flash. User content grows colder over time, so archival photos and posts eventually will move to cold storage systems.

Facebook's overall strategy is to pursue a "disaggregated rack" architecture with designs contributed to the Open Compute Project.

Earlier this year at Open Compute Project Summit, Facebook introduced its "Lightning"- a flexible, NVMe JBOF (just a bunch of flash) box. It is designed to provide a PCIe gen 3 connection from end to end (CPU to SSD). It leverages the existing Open Vault (Knox) SAS JBOD infrastructure to provide a faster time to market, maintain a common look and feel, enable modularity in PCIe switch solutions, and enable flexibility in SSD form factors.

Rao highlighted three other solid storage innovations.

Next gen Non-Volatile Memory (NVM), which promise significant boosts in performance and scalability using 3D technologies and other breakthroughs.

Facebook's own AVA card, which puts four M.2 modules on a card.

WORM, which promises very large capacities (>100TB) with low endurance (150 write cycles) for low-cost, long-term storage.


Intel to Acquire Nervana for AI

Intel agreed to acquire Nervana Systems, a start-up based in San Diego, California, for their work in machine learning. Financial terms were not disclosed.

Nervana, which was founded in 2014, developed a software and hardware stack for deep learning.

Intel said the Nervana technology would help optimize the Intel Math Kernel Library and its integration into industry standard frameworks, advancing the deep learning performance and TCO of the Intel Xeon and Intel Xeon Phi processors.


Nervana is headed by Naveen Rao.

Netronome's 25GbE Adapters Accelerate OpenStack Firewalls and Open vSwitch

Netronome unveiled its the Agilio CX dual-port 25GbE intelligent server adapter (ISA) and Agilio OVS Firewall software integrated with the Mirantis OpenStack solution for seamless cloud-based provisioning.

Netronome said its Agilio 25GbE platform delivers up to 4X higher performance and scale when implementing OpenStack and Linux Firewall-based stateful and zero trust security compared with other 25GbE server adapter solutions in the market.  Agilio also promises up to 5X higher performance for OpenStack and Open vSwitch (OVS)-based acceleration. The Agilio 25GbE platform is capable of delivering up to 25Gb/s of throughput for OVS and Linux Firewall-based OpenStack deployments while consuming only one CPU core. With 2GB of on-board adapter memory, the Agilio platform can support up to two million flows for security policy rules, delivering micro-segmentation-based zero trust security.

“More than 60 percent of OpenStack-based cloud deployments today are based on OVS and Linux Firewall, which are well known to consume too many CPU cycles at 10/25GbE and higher networking speeds, starving VMs and applications,” said Sujal Das, senior vice president and general manager of marketing and strategy at Netronome. “Netronome is the first in the industry to offer a holistic 25GbE platform that not only delivers economies in the area of speeds and feeds but also improves overall server efficiency when running modern and secure data center applications.”

“Zero-trust security implemented using OpenStack security groups fills many holes in traditional perimeter-based firewalling exacerbated by increasing east-west traffic,” said Kamesh Pemmaraju, vice president of product marketing at Mirantis. “The integration of the Agilio CX 25GbE and Linux Firewall platform with Mirantis OpenStack solves this multi-dimensional challenge by improving performance, scale and economies for both east-west traffic and servers implementing zero-trust security.”


Cisco Contributes Full Duplex DOCSIS Chip Design

Cisco is contributing a silicon reference design for Full Duplex DOCSIS to the cable industry on a royalty-free basis.

Specifically, Cisco is offering a validated reference design for a digital echo canceler that maximizes the use of HFC capacity. Over the past 2 year, a team of Cisco engineers have defined a multi-slice scalable echo canceler (EC) for the Full Duplex DOCSIS specification that seamlessly integrates with the Cable Modem Termination System (CMTS) architecture. The technology is scalable for a return path from 200 MHz (1.7 Gbps) to 1.2 GHz (10 Gbps).

“By making this royalty-free design available to the industry, we can help our cable customers evolve to more rapidly deploy virtualized, fiber-deep, and all-IP infrastructures,” said John Chapman, Cisco fellow and CTO, Cable Access Business. “We hope to accelerate the transformation of the cable industry to deliver multi-gigabit speeds and new high bandwidth services and products, and in the near future, customers can begin to enjoy the benefits of Full Duplex DOCSIS technology.”

“DOCSIS 3.1 Full Duplex with up to 50 times more upstream capacity than today and echo cancellation technology, is further evidence that DOCSIS and the cable network itself has a long, useful life ahead,” said Jeff Finkelstein, executive director, Network Strategy, Cox Communications. “The work being done on Full Duplex by the MSO, vendor and CableLabs team shows that the cable network will continue to evolve and deliver high bandwidth services for many years, enabling our customers to enjoy the services most important to them to meet their future needs.”


NEC Debuts 10 Gbps E-band Radio with Adaptive Modulation

NEC Corporation of America launched its new iPASOLINK EX Advanced radio, featuring Adaptive Modulation for 10 Gbps wireless throughput capacity in a single channel.

NEC said its new radio operates in E-band millimeter wave frequency range (71-86GHz), achieves low latency and high transmission capacity of 10Gbps within a single channel, more than three times higher than NEC's previous iPASOLINK EX model.

The new product's Adaptive Modulation and Bandwidth Radio (AMBR) feature enables dynamic adjustment of its modulation method and communication bandwidth according to the particular application, weather conditions and link distance, allowing it to maintain high quality network connections even in severe weather conditions.

The new iPASOLINK EX Advanced is approximately 30% smaller than the previous iPASOLINK EX model, improving site acquisition and installation costs. The integrated 10GbE interfaces allows it to efficiently transmit high capacity data via optical fiber to core networks.

"As data demands have increased, capacity has become a significant concern for network planners who need to ensure throughput, security and availability,” said Matt Engelmann, vice president for NEC Corporation of America. “These concerns apply to all transport mediums and I am pleased to confirm our new radios like iPASOLINK EX Advanced meet these demands and help transport network operators deliver higher quality network services for enhanced user experience. These enhancements are industry leading and indicative of NEC’s continued investment in radio technology.”

Western Digital Achieves First 64 Layer 3D NAND

Western Digital has achieved pilot production of 3D NAND technology in 64 layers.

The technology, which WD is calling BiCS3, was developed jointly with Toshiba, its manufacturing partner. It will be initially deployed in 256 gigabit capacity and will be available in a range of capacities up to half a terabit on a single chip.

WD expects commercial volumes of BiCS3 in the first half of calendar 2017.

"The launch of the next generation 3D NAND technology based on our industry-leading 64 layer architecture reinforces our leadership in NAND flash technology," said Dr. Siva Sivaram, executive vice president, memory technology, Western Digital. "BiCS3 will feature the use of 3-bits-per-cell technology along with advances in high aspect ratio semiconductor processing to deliver higher capacity, superior performance and reliability at an attractive cost. Together with BiCS2, our 3D NAND portfolio has broadened significantly, enhancing our ability to address a full spectrum of customer applications in retail, mobile and data center."