Tuesday, July 11, 2017

Intel Debuts its Xeon Scalable Platform

In what it called its “biggest data center launch in a decade”, Intel officially unveiled its Xeon Scalable platform, a new line of server CPUs based codenamed Skylake and specifically designed for evolving data center and network infrastructure.

The new silicon, which Intel has been refining for the past five years, promises the highest core and system-level performance averaging 1.65x higher performance over the prior generation.  First shipments went out several months ago and are now in commercial use at over 30 customers worldwide, including AT&T, Amazon Web Services and Google.  Intel says every aspect of Xeon has been improved or redesigned: brand new core, cache, on-die interconnects, memory controller and hardware accelerators.

Intel’s new processors scale up to 28 cores and will be offered in four classes: Platinum, Gold, Silver, and Bronze. The design boasts six memory channels versus four memory channels of previous generation for memory-intensive workloads. Up to three Intel Ultra Path Interconnect (Intel UPI) channels provide increase scalability of the platform to as many as eight sockets.

Intel claims 4.2X greater VM capacity than its previous generation and a 65% lower total cost of ownership over a 4-year old server.  Potentially you might need only one quarter of the number of servers. For communication service providers, the claim is that the new Xeon Gold will deliver a 2.7X performance boost for DPDK L3 forwarding applications over a 4-year old server.



Key innovations in Xeon Scalable Platform

  • Intel Mesh on-chip interconnect topology provides direct data paths with lower latency and high bandwidth among additional cores, memory, and I/O controllers. The Mesh architecture, which replaces a previous ring interconnect design, aligns cores, on-chip cache banks, memory controllers, and I/O controllers, which are organized in rows and columns, with wires and switches connecting them at each intersection to allow for turns. Intel said this new design yields improved performance and greater energy efficiency.

    More specifically, in a 28-core Intel Xeon Scalable processor, the Last Level Cache (LLC), six memory channels, and 48 PCIe channels are shared among all the cores, giving access to large resources across the entire die a
  • Intel Advanced Vector Extensions 512 (Intel AVX-512), which delivers ultra-wide vector processing capabilities to boost specific workload performance, now offers double the flops per clock cycle compared to the previous generation.  Intel AVX2,6 Intel AVX-512 boosts performance and throughput for computational tasks such as modeling and simulation, data analytics and machine learning, data compression, visualization, and digital content creation.
  • Intel Omni-Path Architecture (Intel OPA) is the high-bandwidth and low-latency fabric that Intel has been talking about for some time. It optimizes HPC clusters, and is available as an integrated extension for the Intel Xeon Scalable platform. Intel said Omni-Path now scales to tens of thousands of nodes. The processors can also be matched with the new Intel Optane SSDs.
  • Intel QuickAssist Technology (Intel QAT) provides hardware acceleration for compute-intensive workloads, such as cryptography and data compression, by offloading the functions to a specialized logic engine (integrated into the chipset). This frees the processor for other workload operations. Encryption can be applied to data at rest, in-flight, or data in use.  Intel claims that performance is degraded by under 1 percent when encryption is turned on. This function used to be off-chip.
  • Enhanced Intel Run Sure Technology, which aims to reduce server downtime, includes reliability, availability, and serviceability (RAS) features. New capabilities include Local Machine Check Exception based Recovery (or Enhanced Machine Check Architecture Recovery Gen 3) for protecting critical data.

Aiming for the megatrends

In a webcast presentation, Navin Shenoy, Exec Vice President & General Manager, Intel’s Data Center Group, said that as traditional industries turn to technology to reinvent themselves, there are three megatrends that Intel is pursuing: Cloud, AI & Analytics, and 5G.  The new Xeon Scalable Platform addresses the performance, security and agility challenges for each of these megatrends.

AT&T’s John Donovan testifies, performance boost about 30%

During the big Xeon Scalable unveiling, Intel invited AT&T’s John Donovan on stage to talk about the new processors/ AT&T gained access to the new processors a few months ago and has already deployed Xeon Scalable servers which are carrying production traffic.  Donovan reported about at 30% performance boost for its applications over the previous Xeon generation. The net effect he said should be a 25% reduction in the number of servers it will need to deploy.  Intel has been seeding the process with other top customers as well.

This 30% performance boost is certainly good, but it is probably a stretch to call this upgrade “the biggest data center announcement in a decade.” For other applications, perhaps the claim is better justified. One such area is machine learning, which Intel identifies as one of the key megatrends for the industry. There are some interesting developments for Xeons in this domain.

A strong market position

Google Cloud Platform (GCP) is the first public cloud to put the Intel Xeon Scalable Platform into commercial operation. A partnership between Google and Intel was announced earlier this year at a Google event where the companies said they are collaborating in other areas as well, including hybrid cloud orchestration, security, machine and deep learning, and IoT edge-to-cloud solutions. Intel is also a backer of Google’s Tensor Flow and Kubernetes open source initiatives.

In May 2016, Google announced the development of a custom ASIC for Tensor Flow processing. These TPUs are already in service in Google data centres where they "deliver an order of magnitude better-optimized performance per watt for machine learning." For Intel, this poses a long-term strategic threat.  With this announcement, Intel said Xeon’s onboard advanced Vector Extensions 512 (Intel AVX-512) can increase machine learning inference performance by over 100x – a huge boost for AI developers.

The data centre server market is currently dominated by Intel.  Over the years, there have been several attempts by ARM to gain at least a toe-hold of market share in data centre servers, but so far, the impact has been very limited.  AMD recently announced its EPYC processor for data centre servers, but no shipment date has been stated and the current market position is zero. NVIDIA has been gaining traction in AI applications as well as in public cloud acceleration for GPU intensive applications – but these are specialized use cases.

Zayo – Growing by leaps and bound

If anyone is keeping a list of the most acquisitive companies in the telecom space, Zayo Group Holdings must have a spot toward the top. Zayo got its start in 2007, when Dan Caruso, John Scarano and Matt Erickson saw an opportunity for a new play in a game they knew well – building the highest capacity network infrastructure dark fibre and lit services, including carrier-neutral colocation, mobile infrastructure, wavelengths, Ethernet and IP services. All three co-founders had previously worked at Level 3 Communications, which built a reputation for its nationwide, wholesale transport network and competitive local exchange carrier (CLEC) business. Level 3 reached a high-level of prominence by its bold strategy, quick deal making and willingness to raise the huge sums of money needed for a rapid entrance onto the global telecoms stage.  Zayo seems to have taken all of these characteristics and multiplied them 10x.

Zayo by the numbers (currently):

·         122,000 fibre route miles.

·         10 million fibre miles.

·         46 data centres across North America and Europe.

·         Presence in 370 markets.

Quick update on the financial picture

Zayo's revenues are primarily derived by leasing dark fibre (23%), dark fibre for mobile infrastructure (6%), wavelength services (15%), Ethernet services (9%), IP services (7%); interconnect-oriented colo (10.5%), and Allstream Canada (23.1%). For its most recent third quarter, ended March 31, 2017, Zayo reported $550.2 million of consolidated revenue; including $470.9 million from the Communications Infrastructure segments and $79.3 million from the Allstream segment. Net income amounted to $27.0 million, including $17.3 million from the Communications Infrastructure segments and $9.7 million from the Allstream segment.

Getting to this stage

Since its founding in 2007 through to the current quarter, Zayo has been in a constant process of hunting for, acquiring and integrating other network operators. Outside observers give the company especially high marks in network integration, a technically difficult process requiring careful planning and methodical accuracy so as not to disturb live customer traffic on the old or new network. Zayo operates a single platform for delivering a uniform set of services despite having swallowed over 30 networks over the past 10 years. Wikipedia puts the number of Zayo acquisitions at 39. Here they are in reverse chronological order based on the year the deal was first announced:

·         2017 - two data centres in San Diego from KIO Networks for $12 million, located at 12270 World Trade Drive and 9606 Aero Drive, with more than 100,000 sq feet of space and 2 MW of critical, IT power, with additional power available.

·         2016 – Electric Lightwave.

·         2015 - Latisys, IdeaTek Systems, Viatel, a new Dallas, TX data centre from Stream Data Centers, Clearview International and Allstream, Canada's main facilities-based inter-exchange carrier which has its roots in the railway telegraph business.

·         2014 - Dallas-based data centere operator CoreXchange, Neo Telecoms and Geo Networks.

·         2013 - Austin-based data centre operator Core NAP, Access Communications, Midwest-based dark fibre operator FiberLink and Corelink Data Centers.

·         2012 - AboveNet, FiberGate, Arialink, US Carrier Telecom, First Telecom Services, and Maryland-based Litecast/Balticore.

·         2011 - MarquisNet data centre business in Las Vegas and 360networks.

·         2010 - AGL Networks, American Fiber Systems and Dolphinis Cummins Station data centre and colocation services.

·         2009 - FiberNet Telcom Group.

·         2008 - Columbia Fiber Solution, Adesta Communications assets, two sets of fibre assets from Citynet, two sets of fibre assets from Northwest Telephone and CenturyTel Regional Markets.

·         2007 - Memphis Networx, VoicePipe, Onvoy Inc, PPL Telcom and Indiana Fiber Works.

Among the latest acquisitions were two of its most significant buys: Allstream and Electric Lightwave. Allstream was a Canadian leader in IP communications and the only national provider that works exclusively with business customers. Allstream brought over 9,000 route km of metro fibre network concentrated in Canada's top five metropolitan markets (Toronto, Montreal, Vancouver, Ottawa and Calgary) that connect to approximately 3,300 on-net buildings. In addition, Allstream has an approximate 20,000 route km long-haul fibre network connecting all major Canadian markets and 10 U.S. network access points. In addition, Allstream operates colocation space in Toronto, Montreal, and Vancouver. Zayo has previously stated that approximately half of Allstream's revenue is a direct fit with its existing core business. The remaining half will be organised into two additional segments: Voice and Universal Communications (approximately one third of Allstream’s revenue), and Small Business (primarily enterprise voice). Each of these will be separated into standalone business units in parallel with the formation of Zayo Canada. Perhaps Zayo will sell these operations.

Zayo agreed to acquire Electric Lightwave, formerly known as Integra Telecom, for $1.42 billion in cash. The recently-completed acquisition of Electric Lightwave, which provides infrastructure and telecom services primarily in the western U.S., was valued at $1.42 billion. Electric Lightwave had 8,100 route miles of long haul fibre and 4,000 miles of dense metro fibre in Portland, Seattle, Sacramento, San Francisco, San Jose, Salt Lake City, Spokane and Boise, with on-net connectivity to more than 3,100 enterprise buildings and 100 data centres.

Recent management changes

There has been considerable turnover in the company's executive ranks in the past year. Within the year, has Zayo recruited three high-powered execs from Level 3, although one has since gone back to Level 3.

Jack Waters, CTO of Zayo Group, was previously CTO at Level 3 where his responsibilities included global network technology, architecture, engineering, process and security. He had been with Level 3 since 1997.

Andrew Crouch, president and COO, previously served as regional president of EMEA for Level 3, where he also was responsible for its Global Accounts Division.

Ed Morche, president of sales, was previously group VP of sales at Level 3 where he was responsible for sales, sales engineering, business operations, commercial services and metro market expansion for the company's largest business unit – North America’s Enterprise and Government customers. On June 22nd, Morche resigned his new position at Zayo in order to return to Level 3. His decision to return to Level 3 comes just as CenturyLink prepares to complete its acquisition of Level 3.

Telxius deploys Infinera Instant Bandwidth

Infinera announced that Telxius, the global telecommunications infrastructure company formed by Telefónica, which serves around 300 million subscribers worldwide, is offering its customers on-demand, software-defined terabit capacity service activation on its SAM-1 submarine network that links the U.S., Brazil and Puerto Rico.

The new on-demand service capabilities offered by Telxius are enabled by Infinera’s Instant Bandwidth software that is able to activate capacity on the Infinera Intelligent Transport Network deployed by Telxius.

Telxius manages around 65,000 km of international infrastructure encompassing high-capacity optical subsea cable systems, of which 31,000 km are owned by the company. Telxius owns and operates the 25,000 km SAM-1 subsea cable connecting the U.S. with Central and South America, where Infinera solutions are deployed; Infinera solutions are also installed on the terrestrial backhaul routes associated with the cable.

The Telxius Intelligent Transport Network is based on Infinera's DTN-X XTC Series solution with support for up to Tbit/s transmission capacity. The XTC Series is designed to simplify network operations via high-capacity optical super-channels enabled by Infinera's large-scale photonic integrated circuits and integrated packet-aware optical transport network (OTN) switching, which also provides the flexibility to address the requirements of both subsea and terrestrial networks.

In addition, utilising Infinera's Instant Bandwidth software defined capacity (SDC) functionality Telxius is able to deploy bandwidth capacity in 100 Gbit/s increments without the need to pre-deploy additional line cards in the system. Infinera noted that Instant Bandwidth is in use by 60+ customers.


* Earlier this year, Infinera expanded its SDC offering with the introduction of its Instant Network solution for cloud scale networks, which enables service providers to automate optical capacity engineering and quickly scale optical capacity using its Xceed and Digital Network Administrator (DNA) software.

* Telefónica announced in February 2017 that it had agreed the sale of up to 40% of Telxius to KKR for Euro 1,275 million, representing an implied enterprise value of Euro 3,678 million and an equity value of Euro 3,188 million euros for Telxius. Telefónica retained a controlling stake in Telxius.

* Recently, Telxius, Facebook and Microsoft, together with local and regional authorities from Biscay and the Basque Country, announced landing of the 6,600 km trans-Atlantic submarine cable MAREA in Sopelana, Spain as part of a project announced in 2016. MAREA will run between Virginia Beach in the U.S. and Sopelana/Bilbao, Vizcaya in Spain and provide an initial design capacity of 160 Tbit/s. The cable is scheduled to be completed in the autumn of this year.

Cogeco to acquire MetroCast cable systems for $1.4bn

Cogeco Communications, the 8th largest cable operator in North America, announced that its subsidiary serving the U.S., Atlantic Broadband, has entered into a definitive agreement with Harron Communications to purchase all of its cable systems operating under the MetroCast brand.

Under the terms of the agreement, substantially all of the assets of MetroCast will be purchased for $1.4 billion, while in conjunction with the transaction Atlantic Broadband expects to realise tax benefits with a present value of approximately $310 million. After adjusting for tax benefits, the purchase price represents a multiple of approximately 9x 2017 forecast adjusted EBITDA for the business being acquired.

As part of the transaction, Caisse de dépôt et placement du Québec (CDPQ) has committed a $315 million equity investment for a 21% interest in Atlantic Broadband's holding company, with the remaining balance of the purchase price and transaction costs to be financed through a committed secured debt financing provided by two banks at Atlantic Broadband.

MetroCast's networks pass 236,000 homes and businesses in New Hampshire, Maine, Pennsylvania, Maryland and Virginia and serve approximately 120,000 Internet, 76,000 video and 37,000 telephony customers. For the calendar year 2017 the company's revenue is expected to be approximately $230 million, with adjusted EBITDA projected to be approximately $121 million.

Cogeco stated that the transaction is intended to add scale in the American broadband services segment, with Atlantic Broadband's primary service units (PSU) to increase from approximately 602,000 to a total of 835,000 after the acquisition. In addition, MetroCast's systems are mainly serve non-metropolitan markets with attractive demographic profiles and it will be able to leverage Atlantic Broadband's product and sales capabilities to expand the customer base.

The MetroCast network is fully digital and comprises 860 MHz or FTTH infrastructure across 95% of the network and supports broadband speeds of 150 Mbit/s across the footprint. Atlantic Broadband is currently the 9th largest cable company in the U.S. serving around 239,000 Internet, video and voice customers and following the transaction will have a geographic footprint extending from Maine to Florida.

The transaction, which is subject to regulatory approvals and other customary closing conditions, is expected to close in January 2018.

Cogeco Communications is a major cable operator in North America, operating in Canada under the Cogeco Connexion name in Québec and Ontario, and in the U.S. through the Atlantic Broadband brand.


Regarding the transaction, Atlantic Broadband president and CEO Richard Shea said, "After the successful acquisition and integration of the MetroCast Connecticut system almost two years ago, Atlantic Broadband understands the sizable residential and business growth potential that it can expect with the remaining MetroCast systems".


ZTE and China Unicom complete first 5G NR field test in Shenzhen

ZTE announced that it has supported China Unicom as the operator conducted its first 5G NR (New Radio) field test based  on its pre-commercial 5G base station at sub 6 GHz and featuring massive MIMO, LDPC (low-density parity check) and other key 5G technologies.

During the trial China Unicom achieved data rates of up to 2 Gbit/s for single user-equipment. The 5G NR field test carried out in Shenzhen used the 3.5 GHz frequency band with a 100 MHz bandwidth, and was conducted by ZTE working with the Guangdong branch of China Unicom and the operator's network construction department and the China Unicom Network Technology Research Institute.

The field test was designed to verify the 5G technical performance and product commercial capabilities in a live network environment and builds on the established partnership for 5G network development between China Unicom and ZTE.

In 2016, China Unicom launched a 5G laboratory to verify the feasibility of potential key technologies and accelerate the development of 5G base station designs and platforms. Following the start of 5G field testing in 2017, China Unicom is expanding its efforts for verification, with a goal of achieving pre-commercial 5G network deployment in 2019 and a large-scale roll-out in 2020.

ZTE noted that earlier in 2017, it initiated the development of NOMA (non-orthogonal multiple access) technology at 3GPP, representing a core project for 5G NR. ZTE is also involved in China's national 5G tests, having completed testing in seven major scenarios, and was first to complete mMTC field tests in Phase 2 of the program.

In April this year, ZTE announced that it was implementing mobile edge computing (MEC) pilots and technical verification trials with China Unicom, China Telecom and China Mobile, ahead of plans to deploy the technology commercially in 2018. ZTE noted that it launched MEC pilots in collaboration with China Telecom, China Mobile, and China Unicom during 2016.


Huawei details Boundless Computing server strategy and solutions

Huawei announced in Beijing the launch of its Boundless Computing server strategy and series solutions, which focuses on industry requirements for digital transformation and outlines the company's 5-year innovation roadmap for computing.

During the launch event, Huawei delivered an overview of the server strategy and its business positioning, and also launched its V5 series solutions, which include the all-flash SAP HANA appliance solution, a big data application acceleration solution, edge computing for smart video analytics solution, and the G series heterogeneous computing platform.

Huawei's Boundless Computing vision encompasses optimising computing for applications and bringing computing power closer to data sources to help realise the potential of computing. It also includes progressing beyond servers and enabling data centre-level resource pooling and on-demand provisioning to improve the computing efficiency of data centres.

The strategy also involves going beyond the limits the of data centre to enable smart access, and expanding computing into the data sources to enhance the intelligence of data at the remote end.

At the release ceremony, Huawei, together with customers and partners including Industrial and Commercial Bank of China, Tencent, SAP and Microsoft, discussed the challenges to computing development and how they can be addresses. Huawei also released the FusionServer V5 delivering five key smart features.

Describing the strategy, Qiu Long, president, IT server product line at Huawei, said, "A fully connected world is unfolding and computing will be the pivotal force behind everything… the Boundless Computing strategy is about rethinking the road to a fully connected world, that includes unlocking the potential of computing, going beyond the boundary of servers, and extending further beyond the boundary of data centres".


AT&T completes acquisition of Vyatta network OS assets from Brocade

AT&T, which in early June announced it would acquire Vyatta, has completed its the acquisition of the Vyatta network operating system and associated assets of Brocade Communications Systems through an agreement that included the hiring of several dozen Brocade employees, mainly located in California and the UK.
The transaction includes the Vyatta network operating system and vRouter product line. AT&T gains the Vyatta network operating system, including its virtual network functions (VNFs) and distributed services platform, software under development as part of its unreleased product roadmap, existing software licenses and related patents and patent applications.

This acquisition is intended to strengthen AT&T's ability to deliver cloud or premises-based VNFs, beginning with its previously announced SD-WAN cloud service offered with VeloCloud that  was announced in October 2016. The Vyatta acquisition will also help AT&T to enhance its white box platform capabilities.


* In March this year AT&T completed a trial in partnership with other companies and industry groups to design and build white box switches designed to enable more efficient management of data traffic across its network. The trial involved vendors including Barefoot Networks, Broadcom, Delta Electronics, Edgecore Networks, Intel and SnapRoute.

* AT&T stated that the Vyatta platform will help continue to drive its network transformation, through which AT&T is aiming to virtualise and software-control 55% of its network by the end of 2017, rising to 75% by 2020.

LightCounting sees demand for DCI as mega data centres go metro-regional

LightCounting forecasts in its latest Mega Datacenter Optics report that there will be growing demand for high bandwidth interconnectivity as Alibaba, Baidu and Tencent move to a distributed network of data centres in metro areas due to restrictions on building larger facilities in China, while mega data centres operated by the big western cloud companies are transforming into metro-regional clusters.

More specifically, LightCounting notes that Amazon recently disclosed that it operates 25 data centres interconnected with 3,500 fibres in Ashburn, Virginia, while Facebook has expanded its mega data centres by building new facilities near to existing ones, and Microsoft has announced it intends to build more metro and regional data centres to reduce the time involved in planning new facilities.

The research firm states that most of these facilities will be interconnected using DWDM technology, thereby expanding the market opportunity for suppliers of data centre interconnect (DCI) equipment to the cloud companies. Such customers are expected to be early adopters of 200 and 400 Gbit/s DWDM technology.

In terms of vendors, LightCounting notes that Cisco recently introduced the term 'fog computing', referring to adding compute and storage capabilities in facilities located closer to enterprises to facilitate migration of private data centres to the cloud. Meanwhile, Equinix and other colocation providers are increasing their business with enterprise customers as they migrate to the cloud.

In addition, cloud companies are beginning to use colocation data centres to extend reach closer to the end users. LightCounting notes that storage (or caching) of popular videos in local data centres or the central offices of service providers was necessary to meet rising demand for video on demand, while new applications such as SnapChat and Instagram are creating the need to support these applications locally to reduce the load on long haul networks.

The research firm adds that in the future applications such as self-driving cars and IoT will require the use of edge data centres to deliver low latency performance. All of these requirements are driving a transition in the way data centres are deployed and operated.

WorldLink selects Nokia to upgrade backbone in Nepal to 100G

Nokia announced that WorldLink, the largest fixed broadband operator in Nepal, is upgrading a 650-km backbone network using its 1830 PSS (Photonic Service Switch) DWDM technology to support bandwidth-hungry entertainment and enterprise services across the country.

The WorldLink intercity network connects from Kathmandu to Bhairahawa and Birgunj, as well as providing international connectivity between Nepal and neighbouring countries including India. WorldLink serves around 120,000 residential broadband subscribers and 5,000 enterprise broadband circuits, and states it is connecting 10,000 residential FTTH service subscribers each month, creating increasing demand for network capacity.

The deployment of Nokia's optical network technology is designed to enable WorldLink to flexibly increase its network capacity, reach and density leveraging Nokia's programmable chipset, Photonic Service Engine-2 featuring super coherent technology (PSE-2s).

WorldLink is the largest Internet and network service provider in Nepal. The company owns its own leased fibre backbone and purchases IP bandwidth from IP transit locations including Mumbai, Chennai, Singapore and London, interconnecting with service providers such as Tata, Airtel, Singtel and Cogent.

The operator also interconnects with a number of Internet exchanges including NPIX in Nepal, Equinix in Singapore and LINX in London and has partnerships with content delivery network (CDN) operators such as Google, Facebook and Akamai.

WorldLink noted that it operates an advanced network based on equipment from vendors including Juniper and Cisco for IP/MPLS core and aggregation, Extreme Network for switching, Nokia and Huawei for access systems, as well as Nokia DWDM solutions. It operates an extensive national fibre backbone and access network serving 63 districts and offers services including high speed Internet up to 100 Mbit/s and HD IPTV services.


See also