Monday, April 24, 2017

The New Open Source Networking Universe, managed by The Linux Foundation

This month's Open Networking Summit (ONS 2017) in Silicon Valley, which marked the sixth instalment of the annual event since its origin at Stanford University, once again brought many key thought leaders together for keynotes, tutorials, conference sessions, panel discussions and a small exhibition. ONS was the event that set off alarm bells across the industry in 2012 when Google announced that SDN had moved well beyond the research phase and into its commercial network, using a home-grown solution and no routing equipment from conventional vendors.

As far as the current state of open source networking is concerned, perhaps Margaret Chiosi, formerly at AT&T and now with Huawei, put it best, commenting (paraphrased from her keynote), 'open networking has reached the peak of exuberance and not yet crossed the chasm of despair'. In other words, great ideas are flourishing but no one is yet really making money or saving money with open networking. In addition, Guru Parulkar, executive director of ON.Lab, observed that while there are many proof-of-concepts and early deployments, a 'resource gap' of qualified professionals may delay large scale deployments. New technology is difficult and committed resources must be in place if network transformations are to really happen.

ONS 2017 was presented by The Linux Foundation, the San Francisco-based organisation with the lofty ambition of creating the 'largest shared technology investment' in history. A mural at the entrance to the exhibition depicts the New Open Source Networking Universe, and certainly a tremendous amount of intellectual capital is being collected, curated and assembled into specifications and frameworks impacting every layer of the network.

The following is the line-up of networking related projects underway at the Linux Foundation (although there are more non-networking LF projects and other open source networking projects, such as Open Stack, that are not managed by LF).


The Open Network Automation Platform (ONAP) unites two major open networking and orchestration projects, AT&T's open source ECOMP and the Open Orchestrator Project (OPEN-O). This newly launched LF project is one of the largest open source networking initiatives, with members including Amdocs, AT&T, Bell Canada, China Mobile, China Telecom, Cisco, Ericsson, GigaSpaces, Huawei, IBM, Intel, Metaswitch, Nokia, Orange, Tech Mahindra, Reliance Jio, VMware, ZTE and other leading network operators, OEMs and platform providers.


A carrier-grade, integrated, open source platform to accelerate the introduction of new NFV products and services. OPNFV is mainly focused on building NFV Infrastructure (NFVI) and Virtualized Infrastructure Management (VIM) by integrating components from upstream projects such as OpenDaylight, OpenStack, Ceph Storage, KVM, Open vSwitch and Linux. At ONS 2017, OPNFV made its 4th major release, codenamed Danube, adding foundational support for MANO. The next release, codenamed Euphrates and targeted for completion by October, is expected to include more containerisation support.

Open Daylight

One of the largest projects, this promotes and advances the global development, distribution, and adoption of the OpenDaylight (ODL), the largest open source SDN controller. Many industry vendors participate in ODL.


A SDN operating system for service providers that has scalability, high availability, high performance and abstractions to make it easy to create applications and services. At ONS 2017 the ONOS community made its next platform release, adding support for IPv6 routing, vLAN tagged external interfaces and AAA endpoint authentication, a better GUI interface, VPLS support and various southbound enhancements. Notable members of the ONOS community include AT&T, Comcast, China Unicom, Google, NTT Communications, SK Telecom and Verizon.


Central Office Re-architected as a Data Center (CORD) combines NFV, SDN and the elasticity of commodity clouds to bring data centre economics and cloud agility to the telco central office (CO). CORD lets the operator manage central offices using declarative modelling languages for agile, real-time configuration of new customer services. Three versions are underway - Mobile CORD, Residential CORD, Enterprise CORD – and each has multiple proof-of-concept demos.


The Data Plane Development Kit (DPDK) consists of libraries to accelerate packet processing workloads running on a wide variety of CPU architectures. DPDK was created in 2010 by Intel and made available under a permissive open source license. Today, more than 20 key open source projects build on DPDK libraries, including MoonGen, mTCP, Ostinato, Lagopus, Fast Data (, Open vSwitch, OPNFV, and OpenStack. This community has just moved to the Linux Foundation. Gold members of the project are ARM, AT&T, Cavium, Intel, Mellanox, NXP, Red Hat, and ZTE; Silver members include 6WIND, Atomic Rules, Huawei, Spirent, and Wind River. The Korea Advanced Institute of Science and Technology (KAIST), University of Limerick, University of Massachusetts Lowell and Tsinghua University are Associate members.

FRRouting (FRR)

Free range routing (FRR) is a project endeavouring to make the best open source routing stack. FRR, which originated in the Quagga project, includes protocol daemons for BGP, IS-IS, LDP, OSPF, PIM and RIP. At ONS, it was announced that this project will now be managed by the Linux Foundation.

Open Switch (OVS)

An open-source implementation of a distributed virtual multi-layer switch, the main purpose of Open vSwitch is to provide a switching stack for hardware virtualisation in a network.


Platform for Network Data Analytics (PNDA) aims to eliminate complexity by integrating, scaling and managing a set of open data processing technologies and by providing an end-to- end platform for deploying analytics applications and services. The big idea is that open source big data analytics can play in accelerating the transition to more agile, assured and orchestrated services. At ONS 2017, the MEF demonstrated a reference implementation of LSO (Lifecycle Services Orchestration) analytics using PNDA.


AllJoyn is developing an open source framework for IoT.

Cloud Foundry

Cloud Foundry aims to make it faster and easier to build, test, deploy and scale applications.

The Cloud Native Computing Foundation

This group focuses on the development of open source technologies, reference architectures and common formats for cloud-native applications or services.

The Open Container Initiative

Building a vendor-neutral, portable and open specification and runtime for container-based solutions, founding members of this initiative include nine new companies committed to the OCI, with members including: Amazon Web Services, Apcera, Apprenda, AT&T, ClusterHQ, Cisco, CoreOS, Datera, Dell, Docker, EMC, Fujitsu Limited, Goldman Sachs, Google, Hewlett Packard Enterprise, Huawei, IBM, Infoblox, Intel, Joyent, Kismatic, Kyup, Mesosphere, Microsoft, Midokura, Nutanix, Odin, Oracle, Pivotal, Polyverse, Portworx, Rancher Labs, Red Hat,, Scalock, Sysdig, SUSE, Twistlock, Twitter, Univa, Verizon Labs, VMware and Weaveworks.

The Xen Project

This is the leading open source virtualisation platform that powers some of the largest clouds in production today. Amazon Web Services, Aliyun, Rackspace Public Cloud, Verizon Cloud, and many hosting services use Xen Project software.

IBM Joins AIM Photonics Consortium

IBM and The American Institute for Manufacturing Integrated Photonics (AIM Photonics), announced a patent and intellectual property licensing agreement and confirmed IBM has joined the AIM Photonics consortium, which seeks to advance integrated photonic circuit manufacturing and technology development.

Under the intellectual property licensing agreement , IBM will help AIM Photonics establish standard processes in the development of silicon photonics assemblies, such as couplings for communication signals and light sources. Industry and academic AIM Photonics members will access these technologies through a process design kit (PDK) and prototype development at the Rochester Test Assembly and Packaging (TAP) facility and the 300mm chip facility at SUNY Poly's Albany campus.

"Adding IBM as a member of AIM Photonics not only significantly strengthens this outstanding institute, but highlights the momentum the Finger Lakes region is experiencing in the high tech sector," said John Maggiore, New York State photonics board of officers chairman. "This announcement further validates the importance of this institute and the goals it has set out to achieve."

AT&T Fiber Adds 8 new Metros

AT&T has announced that as part of its program to deploy fibre across its service footprint in 21 states it plans to extend its fibre network to parts of eight new metro areas, bringing the total markets where the service is available to 75 major metro areas.

The metro areas where AT&T is planning to roll-out its 100% fibre network are as follows: Dayton (Ohio); Macon (Georgia); Madison (Wisconsin); Monterey-Salinas (California); Savannah (Georgia); South Bend (Indiana); Springfield (Missouri); and western Michigan.

AT&T also announced the launch of fibre-based service delivering up to 1 Gbit/s bandwidth in the East Bay area of California, including in parts of the cities of Fremont, Newark, Oakland and the surrounding areas. The company also expects to offer fibre-based Internet access in parts of Hayward, San Leandro and Union City in California shortly.

AT&T noted that the Oakland metro is one of 52 metro markets nationwide where its fibre Internet service is currently available.

The company stated that it is now offering a 1 gigabit connection over its all-fibre network to more than 4.6 million locations nationwide, and during 2017 plans to add 2 million locations as it progresses towards the target of reaching a total of at least 12.5 million locations with fibre by mid-2019.

  • In mid-February, AT&T Fiber announced it would shortly be launching its gigabit Internet service in the metro areas of Columbia (South Carolina), Jackson (Mississippi). Knoxville (Tennessee), Milwaukee (Wisconsin) and Shreveport, (Louisiana). It stated the expansion would bring the total markets where gigabit Internet service was available to 51.

    At that time, AT&T Fiber also announced the launch of fibre-based gigabit connectivity for customer locations in parts of the Milwaukee area and its expansion in the Louisville area. In Milwaukee AT&T announced the launch of service in parts of Milwaukee, Wauwatosa, Waukesha and surrounding communities; in the Louisville area, it introduced service in parts of Jeffersonville and New Albany in Indiana.

Huawei and GE Partner on Industrial Cloud-based Predictive Maintenance

Huawei and GE Digital announced an Industrial Cloud-based Predictive Maintenance Solution that integrates Huawei's Edge Computing IoT (EC-IoT) with GE's Industrial Internet cloud platform Predix to enable end-to-end connectivity between industrial assets and cloud applications.

The joint solution could be used for real-time machine health monitoring, data analysis and perception, and smart maintenance decision-making. In the digital industrial era, the solution helps manufacturers reduce maintenance costs and prevent unplanned asset downtime, as well as drive innovation in products and services.

"The digital industry era not only brings development opportunities for traditional industries, but also creates challenges with interconnectivity between numerous devices, Big Data computing, and intelligent applications. Huawei collaborates with partners to explore the industrial Internet field and address these challenges based on our experience in digital transformation and leading ICT technology, stated Diana Yuan, President of Marketing and Solution Sales Department of Huawei Enterprise Business Group.

Harel Kodesh, CTO of GE Digital, said: "The growth of the Industrial Internet depends on the cultivation of an ecosystem and a commitment to jointly build solutions to accelerate digital transformation for industry. This new solution brings together the strengths of two great companies – Huawei and GE – and is already driving positive outcomes for innovation-driven customers like Schindler. These partnerships will help industrial companies take advantage of the Industrial Internet and drive unprecedented gains in productivity."

OpenStack User Survey Finds Increasing Deployments

The OpenStack Foundation announced that its user survey for 2016 showed that it achieved 44% more deployments and gained input from 22% more organisations than a year ago, while also revealing increased scale of deployments and a more diverse user base as users seek to avoid vendor lock-in.

OpenStack notes that the ninth user survey indicates that the open infrastructure platform for private and public clouds and telecom networks has growing appeal for organisations of all sizes. Specifically, the survey shows that 32% of users have 10,000 employees or more and 25% of organisations have fewer than 100 employees, with around 61% of users and 74% of deployments physically located outside of the U.S.

In addition, OpenStack notes that the typical (median) OpenStack user is running 61% to 80% of infrastructure on OpenStack, while for larger clouds with 1,000 cores or more, the median user is running 81% to 100% of the overall infrastructure on OpenStack.

Further key findings revealed by the OpenStack user survey include:

1.         The typical size of an OpenStack cloud increased, with 37% of clouds having 1,000 or more cores, compared to 29% in 2015, and 3% of clouds with more than 100,000 cores; the number of users operating Nova cells increased by 218%.

2.         Swift object provisioning also showed increased scale, with 16% of deployments provisioning more than 1 petabyte of object storage (4% last year), while 33% reported storing 100,000 or more objects (13% last year).

3.         The business drivers for OpenStack adoption highlight its competitive advantages, including accelerated innovation and avoiding vendor-lock in; respondents stated that these factors are more important than reducing costs and increasing operational efficiency, which ranked No. 1 and 2 in past surveys.

4.         Containers remain the leading technology of interest to users, with 65% of those running OpenStack services inside containers using Docker runtime, while 47% of those using containers to orchestrate apps on OpenStack employing Kubernetes.

5.         The proportion of OpenStack deployments in production remains stable, with around two-thirds of clouds in production, while number of clouds in proof-of-concept and test stages indicating healthy growth for the future.

6.         A typical deployment runs nine projects, with 16% running 12 or more projects; all of OpenStack's core services are in use by 89–98% of clouds, significantly higher than last year for core projects in production.

Vodafone Turkey Tests GL Spectrum Sharing with Huawei

Vodafone Turkey and Huawei announced they have completed what they claim is the first verification of the GL spectrum sharing solution on Vodafone's commercial 900 MHz networks in the city of Diyarbakir in Turkey.

This solution, developed through a collaboration on spectrum sharing capabilities between Huawei Mobile Innovation Centre and Vodafone Networks CoE (Center of Excellence), enables spectrum sharing between GSM and LTE with a high degree of overlap between the two technologies. The technology improves both LTE data rate and cell capacity available in the 900 MHz spectrum allocation from Vodafone Turkey.

Huawei claims that compared to LTE 5 MHz performance, the GL technology enables average user data throughput to be increased by nearly 58% on the downlink and 44% on the uplink.

Vodafone Turkey launched its 4.5G network for commercial use in April 2016 and claims to offer the largest 4.5G coverage in Turkey, currently serving over 8 million LTE customers. Subsequently, as more users migrate to LTE the effective allocation of spectrum resources to support increasing LTE data traffic has become a priority for Vodafone Turkey. The GL spectrum sharing solution features Huawei's proprietary algorithms, which are designed to free operators from the restrictions of standard LTE bandwidth.

The Huawei solution ensures that scattered spectrum resources are utilised more effectively to enable higher data rates and an enhanced user experience. By providing more flexibility in terms of the resource block allocation and spectrum allocation effectively used by LTE the GL technology also maintains sufficient GSM channels to carry 2G CS traffic.

In February, Huawei, Vodafone and Qualcomm Technologies announced they had jointly created in Turkey what they claimed was the first network ready to use Licensed-Assisted Access (LAA) technology, based on the 3GPP R13 standard. LAA is designed to help improve the user experience by enabling faster download speeds.

The LAA-ready network was tested using a Huawei Lampsite base station in a Vodafone Turkey store in Istanbul and used 40 MHz of unlicensed spectrum in 5 GHz and 15 MHz licensed spectrum in 2.6 GHz for three carrier aggregation. The on-site peak download speed of 370 Mbit/s was achieved using a Qualcomm Snapdragon 835 processor with X16 LTE mobile test device

  • In 2016, Huawei announced that its Active Antenna Unit (AAU) solution was supporting Turkey's deployment of 4.5G networks to enable enhanced mobile broadband (MBB) services. Turkey launched its large-scale plan for the evolution from 3G to 4.5G networks in 2015 and released the first 4.5G spectrum license in August of that year.

Rambus teams with Samsung to offer 56G SerDes PHY

Rambus, a supplier of hardware and software technologies for the communications market, announced that it is partnering with Samsung Electronics to develop its recently launched 56 Gbit/s SerDes PHY on Samsung's 10 nm LPP (low-power plus) process technology.

Rambus' 56 Gbit/s SerDes PHY is designed to deliver enterprise-class performance in challenging signalling environments typical for high-speed communication systems and provides PAM4 and NRZ signalling with a scalable ADC-based architecture. The flexible device architecture targets long-reach backplane requirements to support the transition to 400 Gigabit Ethernet applications.

Rambus unveiled the 56 Gbit/s multi-protocol SerDes PHY, developed on advanced, second generation FinFET process technology, in February this year. The solution is designed to address demanding communications and data centre applications.

The Rambus 56 Gbit/s SerDes PHY features a 28 GS/s ADC integrated directly into the architecture to enable scalability and extended reach. The PHY can support data rates from 9.95 up to 58 Gbit/s across copper and backplane channels with more than 35 dB channel insertion loss for a range of standard interconnect protocols.

The 56G MPS are designed with a system-oriented approach, taking the interface, interconnect and channel into account when optimising performance and features to maximize flexibility in today’s most challenging system environments and applications. This makes the PHYs ideal for many long-reach, copper and backplane enterprise environments.

Also in February, Rambus announced the availability of its High Bandwidth Memory (HBM) Gen2 PHY, developed for GLOBALFOUNDRIES FX-14 ASIC platform. Leveraging the GLOBALFOUNDRIES 14 nm FinFET (14LPP) process technology, the Rambus HBM PHY is designed for networking and data centre applications, in particular systems demanding low latency and high bandwidth memory performance.

Rambus HBM Gen2 PHY is compliant with the JEDEC HBM2 standard and can support data rates up to 2,000 Mbit/s per data pin to deliver an aggregate bandwidth of 256 Gbit/s for data-intensive tasks.

Regarding the collaboration with Samsung, Ryan Lee, VP, foundry marketing team at Samsung Electronics, said, “Network application is one of Samsung's key segments of focus and the composition of its 10LPP network solution and the Rambus 56 Gbit/s solution will… (offer) high performance and low power consumption… for data processing requirements to come".

ZTE and VEON Team on NFVI and vEPC

ZTE and VEON the Amsterdam-based international telecoms company formerly known as VimpleCom, announced they have signed a global framework agreement covering network function virtualisation infrastructure (NFVI) and virtual evolved packet core (vEPC), as a part of which they plan to cooperate of the development of virtualisation technology.

Under the agreement, ZTE becomes a supplier to the VEON group for NFVI solutions, and will deploy large-scale NFVI and vEPC networks in a number of countries where VEON operates, including Russia. The cooperation between the companies is intended to promote VEON's wider NFV strategic planning and digitalisation initiative.

In addition, as part of the partnership ZTE and VEON will continue to collaborate in the field of virtualisation, and ZTE will support VEON as it implements its long-term digital transformation program designed to enhance the services provided to customers.

In early 2016, VEON/VimpelCom announced plans to build a complete virtual network infrastructure to support the delivery of 4G, 3G and 2G mobile data services to customers in five markets, and that it had selected ZTE to build the new network environment.

For the project, ZTE's vEPC solution was to be implemented initially in Kyrgyzstan and Laos, followed by deployments in Uzbekistan, Armenia and Tajikistan, with the roll-out due to be completed during 2016.

Calix Launches Marketing Cloud

Calix, the provider of Subscriber Driven Intelligent Access, announced the release of Calix Marketing Cloud (CMC), a new cloud solution that builds on the company's five years of cloud-based customer support.

Calix noted that marketing teams at service providers may have access to limited, static subscriber data from billing and data consumption, which does not provide insight into subscriber usage and preference trends. The new CMC solution enables better understanding of subscribers by aggregating behavioural data, allowing the marketer to create, target and execute more effective campaigns.

Calix's CMC features behavioural filters that allow the creation of actionable lists of target subscribers extracted directly from back-office systems, which can then loaded into focused marketing campaigns.

The new cloud-based CMC solution delivers key behavioural information to service providers' marketing teams, including:

1.         Subscriber insights: via a nightly data refresh, CMC reveals trends in a single dashboard view, with an emphasis on identifying subscribers that may be interested in a service upgrade and those prone to churn.

2.         Customer segmentation: designed to simplify the segmentation process, providing detailed breakdowns of user types such as gamers, social media enthusiasts and over-the-top (OTT) service users, enabling the creation of contextually relevant product bundles and marketing content and targeted campaigns.

3.         Targeted campaigns: based on subscriber segmentation, CMC allows the export of target groups into automated marketing engines such as Salesforce Pardot or Marketo, for bill inserts, online campaigns via the best social media channel, or campaigns loaded into a CRM system.