Monday, July 10, 2017

Cisco's intent-based, intuitive networking launch – Part 2

An interesting aspect of Cisco's new strategic direction with its Intent-based enterprise networking is that the company has made a major bet on developing proprietary ASICs rather than relying on merchant Ethernet switching silicon from Broadcom, Cavium, Barefoot, Innovium or other merchant semiconductor suppliers. There are a number of reasons why the company may have chosen this approach.

First, with Broadcom dominating the market for Ethernet switching silicon, it does not help Cisco's market positioning when many competing vendors have similar switches with matching port counts and speeds. Cisco commands a margin premium and to defend over the long term it may have been necessary to pursue a silicon development program. Second, the costs of developing switching silicon may have declined in relative terms to the market opportunity. Consider that several start-ups, including Barefoot, Innovium and Xpliant (now part of Cavium), gathered sufficient venture funding to attempt entering the market. If the costs were too high, the angel investors would not have taken the risk, so one can assume that large vendors, such as Cisco do have the financial resources to proceed with in-house silicon development. Third, while merchant silicon is becoming more programmable, Cisco may see benefits in a proprietary chipset with APIs for select ecosystem partners.

Cisco's Intent-based networking builds on its Digital Network Architecture (DNA) and its widely-installed Catalyst line of campus, branch and access Ethernet switches.

The new line of Catalyst switches are the 9000-series -- out of sequence compared to previous Catalyst generations, including the 2900 series, 3600, 3800, 4500, 6500 and 6800 series. The 9000-series branding makes sense, however, when considering the company's data centre switches. The Cisco Nexus 9000 series is its flagship data centre switch, so it makes sense that the Catalyst 9000 series be positioned as the flashing enterprise switch.

Comparison with Nexus data centre switching line

Cisco's Nexus data centre switching line also differentiates from the competition using proprietary silicon but powers some models with Broadcom. The latest generation of Nexus switches, announced in March 2016, includes a new Nexus 9000 model based on its custom ASIC and a Nexus 3000 model based on Broadcom's Tomahawk silicon. In terms of Nexus network architecture and software programmability, Cisco is supporting three choices: its own, full-bore Application Centric Infrastructure (ACI) architecture, running on the APIC controller in its Nexus 7K and flagship 9K switches, a programmable fabric vision that could also be supported on the new Nexus 3000 switches with Broadcom silicon, and a lighter programmable network architecture running on any of the Nexus switches and featuring NX-OS enhancements for devops, automation and segment routing.

Performance is a key argument for the custom silicon. With the new Nexus 9000 switches, Cisco said it will achieve industry-leading performance for 100 Gbit/s, with 25% more non-blocking performance, at 50% the cost of comparable solutions, plus greater reliability and lower power. In addition, Cisco is announcing a new Nexus Fabric Manager that automates the fabric lifecycle management with a point-and-click web interface and offers automated configuration snapshots and rollback. Nexus Fabric Manager builds and self-manages a VXLAN-based fabric, dynamically configuring switches based on simplified user-based actions. An IT manager can fully deploy a VXLAN-based fabric in just three steps, complete with zero touch provisioning, and can upgrade all fabric switches to a new software release in 'four mouse clicks'.

There are no performance comparisons yet for the Unified Access Data Plane (UADP) 2.0 ASIC that powers the new Catalyst 9000 series, but the company has said that it is highly programmable. One example of this flexibility is the unique encrypted traffic analytics that will supported. This capability leverages the power of the ASIC along with Cisco's Talos cyber intelligence and machine learning to analyse metadata traffic patterns. Cisco said this enables the network to identify the fingerprints of known threats even in encrypted traffic, without decrypting it and impacting data privacy. Cisco claims it can detect threats in encrypted traffic with up to 99% accuracy, with less than 0.01% false positives. This capability was developed essentially over the course of several months.

Automation and programmability key themes for switching silicon

Other silicon vendors are pursuing automation and programmability as well. Earlier this month, Broadcom unveiled a new generation of its widely-deployed Trident switching silicon for data centre, enterprise and service provider networks. The new StrataXGS Trident 3 switch family, aimed at networks transitioning to high density 10/25/100 Gigabit Ethernet, is manufactured in 16 nm and designed to support fully programmable packet processing, while achieving significant cost and power efficiency. It builds on Broadcom's widely deployed StrataXGS Trident and Tomahawk switch products, offering fully programmable, line-rate switching. It supports new protocol parsing, processing and editing for Service Function Chaining, network virtualisation and SDN. It offers programmable support for new switch instrumentation capabilities such as in-band and out-of-band network telemetry. The StrataXGS Trident 3 also retains complete functional compatibility to with StrataXGS Trident 2 and Trident 2+ based networks, widely adopted by network equipment manufacturers. The chip supports forwarding databases for L2 switching, L3 routing, label switching and overlay forwarding. Broadcom promises 3xincreased ACL scale to support evolving policy/security requirements.

Barefoot Networks, a start-up based in Palo Alto, California, has garnered considerable attention for its Tofino switching chip which boasts 6.5 Tbit/s of overall capacity, and a strategic investment from Google, as well as engagements with  Alibaba, Baidu and Tencent . Barefoot has said its Tofino silicon, which has been sampling since November 2016, excels for its programmable forwarding plane. Recently, the company disclosed ongoing work with AT&T and SnapRoute to deliver what it believes is the first real-time path and latency visualisation. Utilising Tofino and In-band Network Telemetry (INT), AT&T was able to gain deep insight into the network down to packet-level for the first time to help to address bottlenecks caused by path or latency variation. Barefoot noted it took 6 weeks to develop the visualisation capability before it was deployed into AT&T's production environment carrying live customer traffic over a Washington DC to San Francisco link. Barefoot has also announced a partnership with Taiwan-based network switching equipment firms Edgecore Networks and WNC. As a side note, Barefoot Networks was co-founded by Nick McKeown, a Stanford professor and co-founder of Nicira (acquired by VMware), Martin Izzard, Pat Bosshart and Dan Lenoski VP Engineering.

Inovium is another Silicon Valley based start-up that is making a run at scalable Ethernet silicon for data centres switches. Its TERALYNX is claimed to be the first single switching chip to break the 10 Tbit/s performance barrier. It also offers telemetry, line-rate programmability, the largest on-chip buffers and ultra low-latency. The chip is expected to sample in Q3 2017. Innovium’s TERALYNX promises to include support for 10/25/40/50/100/200/400 Gigabit Ethernet standards. It will deliver 128 ports of 100 Gbit/s, 64 ports of 200 Gbit/s or 32 ports of 400 Gbit/s in a single device. This may make it more suited for data centre switches than campus or branch offices.

Some concluding observations

While Cisco may be going down the path of proprietary silicon for its flagship switching platforms, the many other players in this market segment appear set to benefit from innovations coming from Broadcom and the new merchant silicon suppliers. Arista, in particular, continues to thrive despite the many legal challenges presented by Cisco. Earlier this month, Arista rolled out key enhancements to its R-Series platforms, which are based on Broadcom's Jericho+ switching silicon. The latest Arista R-Series platforms provide over 150 Tbit/s of capacity for switching and routing applications with cloud-driven Arista EOS software technologies including Arista FlexRoute and AlgoMatch. Arista is also known to use Cavium's programmable Xpliant switching silicon in other platforms, giving it a multi-vendor approach to the market.


AT&T partners with Coral to launch $200m VC fund

AT&T announced it is committing to invest up to $200 million in a venture capital fund as part of its ongoing effort to develop solutions to address current and emerging technology challenges.

Under the program, AT&T will work with venture capital firm Coral's Communications Industry Platform (CIP) team to identify and invest in start-up companies that are focused on developing technology for connected services and platforms.  AT&T and Coral will also seek to identify additional companies interested in investing in the fund.

The new fund will specifically invest in technologies that run on the Open Network Automation Platform (ONAP) operating system for software-defined networks. AT&T noted that ONAP was created through the merging of a platform developed by AT&T Labs and an existing open source project.

ONAP is currently being used to manage the company's own cloud network, and is now an open source platform hosted by the Linux Foundation that is increasingly being adopted as the standard for virtualised networks worldwide.

AT&T stated that this latest investment initiative builds on its existing innovation programs that encompass AT&T Labs and the AT&T Foundry innovation centre that were established in 2011 to support work with the start-up companies and the open source community.

Coral Group is a venture capital firm that specialises in telecom systems integration, Internet-based consumer and enterprise-facing applications and services. The Communications Industry Platform (CIP) combines venture capital and integrated solutions to help communication service providers address strategic challenges. Coral has provided VC funding to companies including Calix, Flexlight Networks, Infinera and Picolight.


Regarding the initiative, Andre Fuetsch, CTO and president of AT&T Labs, noted, "This investment is part of AT&T's push to address the needs of global service providers… it will collaborate with Coral and other CIP members to find, and even create, start-up companies to build disruptive technologies to solve these challenges".


Baidu deploys Xilinx FPGAs for cloud acceleration

Xilinx announced that Baidu has deployed Xilinx FPGA-based application acceleration services into its public cloud, specifically for the Baidu FPGA Cloud Server, a new service that leverages Xilinx Kintex FPGAs, tools and the software required for hardware-accelerated data centre applications such as machine learning and data security.

The Baidu FPGA Cloud Server provides a complete FPGA-based hardware and software development environment, including hardware and software design examples, and is designed to help users quickly develop and migrate applications with reduced development costs.

The Baidu service is based on each FPGA instance serving as a dedicated acceleration platform that is not shared between instances or users. The design examples provided services including cover deep learning acceleration, encryption and decryption.

Xilinx claims that FPGA-enabled servers can deliver a 10x to 80x performance per watt advantage compared to CPU-only servers. In addition, as they are dynamically reconfigurable, Xilinx FPGAs can support a range of workloads, including machine learning, data analytics, security and video processing.



  • Separately, Baidu announced a partnership with Microsoft for its new open source autonomous driving platform, Apollo. Baidu unveiled Apollo in April, featuring cloud services, software and reference hardware/vehicle platforms, and expects the technology will be running on roads by late 2020.
  • In addition, Conexant, a provider of audio and voice technology solutions, announced it was collaborating with Baidu to release development kits and reference designs for device makers to develop far-field voice-enabled artificial intelligent (AI) devices running on Baidu's DuerOS platform. The development kits and reference designs will feature Conexant's CX20924 4-microphone and CX20921 2-microphone voice input processing solutions and DuerOS, a conversation-based AI system that enables access to a voice-activated digital assistant for mobile phones, TVs and other devices.

Equinix provides direct access to Oracle Cloud in Sydney

Equinix, the global interconnection and data centre company and a Gold level member of the Oracle Partner Network (OPN), announced the immediate availability of dedicated, private access to Oracle Cloud in its Sydney, Australia, International Business Exchange (IBX) data centre.

Available via Oracle Cloud Network Service - FastConnect and the Equinix Cloud Exchange, access will be available for Oracle Infrastructure as a Service (IaaS) as well as Platform as a Service (PaaS). Direct access enables enterprise customers in the region to migrate compute, applications and data to Oracle Cloud in a low-latency manner for an optimal user experience.

The latest expansion builds on previous announcements between Equinix and Oracle to offer direct connection to multiple Oracle PaaS and IaaS services, including database, Java, integration, analytics, compute and storage, in multiple regions worldwide. The addition of Sydney brings the total markets where Equinix is offering private access to Oracle Cloud to five.

In addition to direct access to the Oracle Cloud, customers in Sydney also have access to Oracle Managed Cloud Services to help them determine the best deployment model for their business needs. Oracle Cloud delivers nearly 1,000 SaaS applications and 50 enterprise-class PaaS and IaaS services to customers in more than 195 countries.

Equinix noted that as business models become interdependent and enterprises adopt an interconnection oriented architecture (IOA) strategy in order to deliver the performance users require while utilising multiple IaaS, PaaS and SaaS cloud services. Enterprises can bring these services closer to end users by putting workloads and application near to the digital edge of their network in Equinix facilities, colocating IT hubs adjacent to the cloud service providers that host their cloud edges in Equinix facilities.

Equinix stated that its data centres in Sydney are the most interconnected in Australia, with customers able to establish direct links to both of the continent's largest peering points, as well as key submarine cable systems, and gain direct access to multiple network and cloud providers such as Oracle via the Equinix Cloud Exchange.


The Equinix Cloud Exchange is currently available in 21 markets, namely Amsterdam, Atlanta, Chicago, Dallas, Frankfurt, Hong Kong, London, Los Angeles, Melbourne, New York, Osaka, Paris, Sao Paulo, Seattle, Silicon Valley, Singapore, Sydney, Tokyo, Toronto, Washington DC and Zurich.


Huawei Marine selected to build 4,000km Western Pacific Cable

Huawei Marine, the joint venture between Huawei Technologies and UK-based Global Marine Systems, announced it has signed a contract with the Solomon Island Submarine Cable Company (SISCC) to construct the first submarine cable in the Solomon Islands.

Huawei noted that the Solomon Islands encompasses six major islands and over nine hundred smaller islands situated to the northeast of Australia. Due to its geographical location, the Solomon Islands has been dependent on satellite connectivity for telecommunications service, which is costly and is becoming insufficient to meet the bandwidth demands of the country.

In 2016, the Solomon Islands' government founded the SISCC, with the remit to develop a submarine cable system to connect the main islands and provide onward connectivity to Australia.

Huawei Marine, in conjunction with its parent Huawei Technologies, has been selected to design and construct a network that will incorporate approximately 4,000 km of submarine cable and deliver total capacity of 2.5 Tbit/s. The new cable system will link Sydney on the east coast of Australia to the Solomon Islands capital Honiara, with a further domestic connectivity on to the cities of Auki and Noro. The cable system is expected to be ready for service in 2018.

The signing ceremony for the cable system contract was attended by Solomon Islands Prime minister Manasseh Sogavare, Minister of Finance and Treasury Snyder Rini, Minister of Communications and Aviation Peter Shanel, plus the CEO of SISCC Keir Preedy and the president of Huawei's South Pacific Region, Wei Chengmin.

Huawei Marine noted that it was recently selected to build submarine infrastructure in Papua New Guinea.

Commenting on the project, Keir Preedy, CEO of SISCC, said, I have been planning this submarine cable for nearly seven years… the completion of this cable will solve problems we are facing of insufficient bandwidth, high cost and unstable services… it will provide quality telecommunication services to carriers in the South Pacific region, and deliver high-speed Internet and telecom services in the Solomon Islands".


Fujitsu expands RunMyProcess PaaS

Fujitsu announced the global extension of its digital platform-as-a-service, RunMyProcess, with the addition of new regional cloud platforms in North America and Australia, expanding in the existing platforms in Europe and Asia.

The capacity increase for RunMyProcess is designed to support and address the growing digital transformation needs of organisations with regional compliance requirements for data sovereignty, regulation, security and low-latency.

Fujitsu's RunMyProcess is a native cloud platform that allows organisations to quickly and securely build, test, deploy and scale device-independent applications that connect digital business processes across cloud, on-premises and mobile environments.

Using RunMyProcess, organisations can model and streamline complex business processes while also integrating existing systems and services, whether cloud services such as Office 365 or on-premises solutions such as SAP, with smartphones, tablets, PCs, wearable technology and other devices. The platform's pre-built connectors is designed to allow customers to develop connected applications faster and deploy them within days, as well as scale up to thousands of users.

RunMyProcess is available globally with a choice of regional data centres. The new deployments in Australia and North America are now available both directly from RunMyProcess and via Fujitsu local sales teams.

Fujitsu stated that RunMyProcess is currently serving over 500 customers worldwide across a range of industries including government, defence, retail and manufacturing and utilities, as well as financial services and healthcare.

Fujitsu cited HomeServe USA, a provider of home emergency repair services, as an example of how RunMyProcess was used to digitalise end-to-end processes for the introduction of new products. RunMyProcess worked with HomerServer USA to create a new central process flow, supported by seven sub-processes and connecting 21 cross-functional teams.


Bahamas-based Cloud Carib builds data centres

Bahamas-based Cloud Carib, a provider of enterprise-grade, private and hybrid cloud solutions, announced that it is extending its reach across the Caribbean and Latin America region via multiple new data centres.

Cloud Carib is a leading managed cloud service provider in the Caribbean. The company currently operates data centres located on the islands of Nassau and Freeport in the Bahamas, and plans to launch a newly rebuilt CaribPod data centre in Panama. Cloud Carib has also extended services into Barbados and Jamaica, and is planning to establish additional regional sites in Trinidad, Cayman and other locations before the end of the year.

Cloud Carib data centres are supported 24/7 by the company's command and control centre to provide clients with support and monitoring services. Cloud Carib data centres are operated in accordance with international standards, with services supported by the Cloud Carib service management framework.

The company noted that due to the geographical advantages of the region, there is growing interest from organisations in cloud services provided throughout the Bahamas and Caribbean. Cloud Carib is aiming to address this growing demand through regional expansion as it seeks to become a the premiere managed cloud services provider in the region.

Cloud Carib focuses exclusively on providing managed cloud services leveraging core competencies across disciplines including data centre services, business continuity services and mobile and productivity services.


Headquartered in Nassau, the Bahamas, Cloud Carib offers multiple data centre locations across the Caribbean, including in Freeport, Nassau, Panama, Barbados and Jamaica. It offers solutions including: IaaS, security, business continuity, productivity and mobility, together with professional services and VAR options.


Accedian appoints Dion Joannou as COO

Accedian, the specialist provider of performance assurance solutions for mobile networks and enterprise-to-data centre connectivity, announced the appointment of Dion Joannou as chief operating officer (COO), with responsibility for overseeing the company's global sales, business development, marketing and operations.

As COO, Dion will draw on extensive industry experience and knowledge to support Accedian as it seeks to scale its established market position and expand its customer base in the network operator, service provider and enterprise sectors worldwide.

Dion Joannou joins Accedian from Viavi Solutions (formerly JDSU), where he served for two years as SVP of Sales for the company's Network Service Enablement (NSE) unit. Prior to that, Mr. Joannou was with Nortel Networks for 14 years, where he held a number of leadership positions in the IP and wireless businesses in Europe and Central and Latin America. While at Nortel, he served in roles including chief strategy officer and as president of North American operations.

Dion Joannou holds an MBA in international business from the University of Miami, and a BA in business administration from Southern Illinois University.



  • Accedian recently announced the appointments to its board, namely Steve Pusey, former group CTO of Vodafone Group, Steve Mills, former VP for IBM Software and Systems, and Peter Griffiths, former EVP at CA Technologies. The board appointments followed the acquisition of the company by Bridge Growth Partners in March 2017, and are intended to support company growth.