Friday, March 25, 2022

FCC adds Kaspersky, China Telecom, China Mobile to Security List

The FCC's Public Safety and Homeland Security Bureau added AO Kaspersky Lab, China Telecom (Americas) Corp, and China Mobile International USA Inc. – to its list of communications equipment and services that have been deemed a threat to national security, consistent with requirements in the Secure and Trusted Communications Networks Act of 2019.

“Last year, for the first time, the FCC published a list of communications equipment and services that pose an unacceptable risk to national security, and we have been working closely with our national security partners to review and update this list,” said Chairwoman Jessica Rosenworcel.  “Today’s action is the latest in the FCC’s ongoing efforts, as part of the greater whole-of-government approach, to strengthen America’s communications networks against national security threats, including examining the foreign ownership of telecommunications companies providing service in the United States and revoking the authorization to operate where necessary.  Our work in this area continues.” 

Thursday, March 24, 2022

Havhingsten Cable System links Ireland, UK and the Nordics

Construction has completed on the Havhingsten cable system linking Ireland, the UK and the Nordics. Partners in the project include Aqua Comms, Bulk Fiber Networks, a leading builder and operator of fiber network infrastructure, and Meta. The subsea cable was built by Alcatel Submarine Networks (ASN).

The Havhingsten cable system runs across the Irish Sea from north of Dublin, Ireland to south of Blackpool on the west coast of the UK, with a section also linking into the Isle of Man, and from Seaton Sluice in Northumberland to Denmark. Notably, the route avoids congested paths across the English Channel and ensures lower latency between Dublin and the Nordic region.

The partners are highlighting several key innovations and technological milestones:

  • Havhingsten is the world’s first aluminum conductor powered subsea cable system. As a material, aluminum allows for a much lower cable conductor voltage drop, which ultimately allows for a higher number of fiber pairs per cable. Using aluminum instead of copper reduces cost and weight, allowing more cable to be loaded onto an installation vessel. It also offer improved resistance to hydrogen penetration.
  • The end-to-end system combines data transmission seamlessly across both an unrepeatered subsea segment in the Irish Sea, a terrestrial segment in the UK and a repeatered segment in the North Sea. Typical systems have one or two of these elements, but not all three.
  • The system utilized a new enhanced, jet-assisted burial plough in both the North and Irish Sea segments, allowing the installers of the system to bury the cable to our specific demanding level of protection in very challenging seabed conditions along the route.

https://www.facebook.com/search/top/?q=Havhingsten

Intel and Microsoft contribute Scalable I/O Virtualization spec to OCP

 Intel and Microsoft have contributed a Scalable I/O Virtualization (SIOV) specification to the Open Compute Project (OCP), enabling virtualization of PCI Express and Compute Express Link devices in cloud servers. 

SIOV is hardware-assisted I/O virtualization with the potential to support thousands of virtualized workloads per server. SIOV moves the non-performance-critical virtualization and management logic off the PCIe device and into the virtualization stack. It also uses a new scalable identifier on the device, called the PCIe Process Address Space ID, to address the workloads’ memory.

SIOV technology is supported in the upcoming Intel Xeon® Scalable processor, code-named Sapphire Rapids, as well as Intel Ethernet 800-series network controllers and future PCIe and Compute Express Link (CXL) devices and accelerators. Linux kernel upstreaming is underway with anticipated integration later in 2022.

“Microsoft has long collaborated with silicon partners on standards as system architecture and ecosystems evolve. The Scalable I/O Virtualization specification represents the latest of our hardware open standards contributions together with Intel, such as PCI Express, Compute Express Link and UEFI,” said Zaid Kahn, GM for Cloud and AI Advanced Architectures at Microsoft. “Through this collaboration with Intel and OCP, we hope to promote wide adoption of SIOV among silicon vendors, device vendors, and IP providers, and we welcome the opportunity to collaborate more broadly across the ecosystem to evolve this standard as cloud infrastructure requirements grow and change.”

The first I/O virtualization specification, Single-Root I/O Virtualization (SR-IOV), was released more than a decade ago and conceived for the virtualized environments of that era, generally fewer than 20 virtualized workloads per server.

https://www.intel.com/content/www/us/en/newsroom/opinion/new-specification-hyperscale-virtualization.html

https://www.opencompute.org/

Australia's NBN pushes for FTTP upgrades

NBN Co is on track to enable up to 8 million premises across Australia to access nbn Home Ultrafast, offering peak download speeds of 500 Mbps to close to 1 Gbps, by the end of next year. 

More than 4.4 million premises connected to the network can already access nbn Home Ultrafast, with 2.4 million premises able to access wholesale download speeds of 500 Mbps to close to 1 Gbps via Hybrid Fibre Coaxial (HFC) and around 2 million premises able to access wholesale download speeds of close to 1 Gbps via Fibre to the Premises (FTTP). 

From today, NBN Co is inviting more than 50,000 customers currently served by Fibre to the Node (FTTN) to upgrade to Fibre to the Premises (FTTP). 

To trigger a full fibre upgrade, eligible customers will need to place an order with a participating retailer for a plan based on an eligible wholesale speed tier. These include nbn’s three highest speed tiers: nbn® Home Fast, which offers wholesale download speeds of up to 100 Mbps; nbn® Home Superfast, which offers wholesale download speeds of up to 250 Mbps, or nbn® Home Ultrafast (FTTP)1,2, which offers wholesale download speeds of close to 1 Gbps3,4. 

As part of the company’s $4.5 billion network upgrade program, nbn has so far announced the suburbs and towns across Australia where up to 1.6 million premises currently served by FTTN will become eligible to upgrade to FTTP. 

In addition, The Australian Government will invest $480 million in the nbn® Fixed Wireless network. NBN Co will commit an additional $270 million to the program designed to deliver faster wholesale speeds2 for regional Australia, benefiting thousands of homes and businesses. 

 NBN Co will use the funding to 5G-enable its network of more than 2,200 Fixed Wireless infrastructure sites and more than 22,000 cells in semi-rural areas and across regional and remote Australia. 

https://www.nbnco.com.au/corporate-information/media-centre

Nokia to supply optical transport for TIME in Malaysia

Nokia has been selected by TIME dotCom (TIME) to deploy an optical network linking data centers and cable landing stations in Malaysia. 

The deployment leverages Nokia’s 1830 PSS platform, PSE-V chipset and 400G multi-haul pluggable optics. The solutions will also offer Nokia next generation iROADM line system capabilities with flexible grid, contentionless, directionless and colorless (CDC-F) optical add-drop and restoration features, which eliminates the need for on-site visits to change wavelength connectivity and simplifies network planning. Built-in Layer 0 GMPLS capability, combined with CDC-F, enables automated traffic restoration and network re-optimization for maximum network uptime and capacity utilization.

https://www.nokia.com/about-us/news/releases/2022/03/24/nokia-selected-by-time-dotcom-to-build-high-resilience-cross-peninsular-optical-network-in-malaysia/

Intel updates power supply specs for PCIe 5.0

Intel has published updated ATX 3.0 power supply specifications for next-generation hardware and upcoming components built for technologies like PCIe Gen 5.0.  

The initial ATX 2.0 specs were introduced in 2003. 

Intel said a new 12VHPWR connector will power most, if not all, future PCIe 5.0 desktop Add-in cards (e.g., graphics cards). This new connector provides up to 600 watts directly to any PCIe 5.0 Add-in/graphics card. It also includes sideband signals that will allow the power supply to communicate the power limit it can provide to any PCIe 5.0 graphic card.

New guidelines reflect the PCIe CEM Gen 5 power excursion limit for PCIe 5.0 add-in cards that was published in November 2021. Updated specifications include new DC output voltage regulation that will be necessary for managing new power excursion requirements.

Intel has also revised its ATX12VO spec to provide an updated blueprint for designing power supply units (PSUs) and motherboards that reduce power draw at idle

“Power supplies based on ATX 3.0 and ATX12VO 2.0 will ensure anyone looking to get the most stable and cost optimized performance possible with highest power efficiency out of their desktop PCs will be able to do so – both now and in the future,”states Stephen Eastman, Intel platform power specialist.

https://www.intel.com/content/www/us/en/newsroom/news/intel-introduces-new-atx-psu-specifications.html

Canada's TELUS ships ZTE's Connect-Hub 5G Gateway

TELUS has begun offering ZTE's Connect-Hub 5G Internet gateway for residential fixed wireless service in Canada.

The Connect-Hub 5G Indoor and Outdoor Units support 4G LTE, 5G SA and NSA modes, as well as Sub-6GHz. Advanced antenna technology and algorithms independently select the strongest network signal. The Indoor Unit supports up-to-30 simultaneous Wi-Fi connections and features an innovative antenna that creates a 360-degree dual-band Wi-Fi coverage area.

“Wireless 5G will provide next level speed and connectivity for our wireless high-speed internet customer segments, especially those in rural areas,” said Dwayne Benefield, Senior Vice President, Connected Home & Entertainment at TELUS. “TELUS selected the Connect-Hub 5G™ for its ability to deliver reliable high-speed internet with a simple install for our customers. It supports all our current spectrum bands, in particular our newly acquired 3500 MHz, ensuring customers will be able to reliably stream, conference and game at higher speeds than ever before.”

Deutsche Telekom ends software development in Russia

Deutsche Telekom confirmed that it is closing it software development activities.

In a follow-up to its press conference on February 24, the German carrier issued the following statement:

"Deutsche Telekom does not operate any networks in Russia. We have no business relationships with companies in Russia. However, we have a team of software developers, mainly in St. Petersburg, who provide services to customers outside Russia.

In recent weeks, we have given these employees the opportunity to work outside Russia. Many employees have taken that opportunity and have left the country. We have also ensured that we maintain services for our international customers as best we can without the Russia site."

https://www.telekom.com/en/company/details/deutsche-telekom-ends-its-software-development-activities-in-russia-1001920


Wednesday, March 23, 2022

Oracle Cloud Infrastructure continues to build

Oracle Cloud Infrastructure (OCI) is expanding with 11 new compute, networking, and storage services and capabilities that enable customers to run their workloads faster and more securely at lower costs. 

“OCI continues to break the rules in the cloud, helping customers run their workloads faster, more securely, and more economically,” said Clay Magouyrk, executive vice president, Oracle Cloud Infrastructure. “Customers can build cloud native apps on OCI with support for open, standards-based Kubernetes, while AI and high-performance computing customers can build some of the fastest computing clusters in the cloud.”

Newly announced services:

Container Instances: Enables customers to use containers without directly managing the hosting VM or requiring Kubernetes orchestration. OCI takes care of creating the instance with a secure OS image, networking and storage.

AMD E4.Dense Compute Instances: Enables customer workloads that benefit from attached NVMe drives that provide low-latency storage. This includes database workloads (e.g., relational databases, NoSQL databases), virtualized direct-attached storage, caching, and data warehousing.

Oracle Cloud VMware Solution on AMD: Provides customers with new AMD-based 32, 64 and 128 core options, providing them with industry-leading VM deployment density options per SDDC host, which cater to high CPU or high memory use cases. OCI delivers over 2.5x the memory and CPUs per host than other offerings.

Content Delivery Network (CDN) Interconnect: Establishes direct peering connections with select third-party CDN providers to offer no cost outbound bandwidth for OCI Object Storage. Origin-to-CDN costs are a significant portion of overall CDN costs and OCI is dramatically reducing those costs for customers, even for third-party CDNs. OCI currently offers this capability in North America for Cloudflare CDN.

Content Delivery Network (CDN) Service: Enables customers to deliver digital content to end users from a nearby location in a geographically distributed network. Customers benefit from integrated APIs, console, UCM billing and stronger integrations between OCI Object Storage, Compute and CDN to reduce origin-server-to-CDN egress fees.

Flexible Web Application Firewall (WAF): Enables customers to define a single WAF policy to protect applications from common exploits (e.g., OWASP Top 10) and enforce the policy on load balancer or on the edge.

Web Application Acceleration (WAA): Supports caching and compression of web HTTP responses in load balancer.

Network Visualizer: Allows customers to perform a configuration-based connectivity check and visualize the network path(s) along with information about the virtual network entities in the path. This helps customers identify and fix common virtual network misconfigurations.

vTAP: Enables OCI Network packet capture and inspection out of band to facilitate troubleshooting, security analysis, and data monitoring without impacting performance.

Flexible Block Volumes with Performance-based Auto-tuning: Enables customers to change the performance characteristics of block storage volumes automatically in response to fluctuating demand. This is a unique capability in the cloud market today and helps customers meet peak demands automatically, and helps reduce storage costs when demand is low.

High Availability ZFS: Packages the ZFS file server in a highly available, automated deployment stack that uses OCI Block Volumes for the underlying raw storage.

https://www.oracle.com/news/announcement/oracle-launches-new-flexible-cloud-infrastructure-services-2022-03-15/

Oracle Cloud announces major expansion - 14 new regions worldwide

Citing triple-digit growth, Oracle announed plans for a major expasion of its infrastructure worldwide.Over the next year, Oracle plans to open 14 cloud regions with new locations across Europe, the Middle East, Asia Pacific, and Latin America. Upcoming cloud regions include Milan (Italy), Stockholm (Sweden), Marseille (France), Spain, Singapore (Singapore), Johannesburg (South Africa), Jerusalem (Israel), Mexico, and Colombia. Additional second...


Yamaichi and JAE develop 112Gbps PAM4 cable

Yamaichi Electronics Co. and Japan Aviation Electronics Industry have jointly developed a prototype, high-speed jumper cable interconnect product, which supports the 112Gbps PAM4.

Yamaichi is a supplier of of high speed connectors for data center networking. JAE offers connector and cable harness products for the wide range of market segments including 5G data networking market. 

This interconnect solution has been designed with a compact connector size to be able to fit in a high density design requirement, and can be utilized as a direct jumper cabling from the front panel pluggable IO to host chip, or the internal chip to chip connection.





Tarana raises $170M for fixed wireless in unlicensed spectrum

Tarana Wireless, a start-up based in Milpitas, California, announced $170 million in new funding at a $1 billion valuation. The round was co-led by Axon Capital, Khosla Ventures, and Prime Movers Lab. Also participating were Michigan State Pension Fund, I Squared Capital, and two other large, industry-leading institutional investors.

Tarana also announced that after commercial launch in late 2021 of its G1 platform, the company is on track to deliver over $100 million of revenue in 2022, from a customer base that has exploded to more than 120 service providers. Tarana also announced it has raised $170 million 

“The key ingredient driving our rapid expansion is overwhelming customer demand for G1, the world’s most advanced fixed wireless access (FWA) platform,” noted Tarana CEO Basil Alwan. “Like fiber, G1 delivers ultra-fast internet service with impressive upstream and downstream speeds, but with dramatic improvements in both cost and time to deploy. Finally, true high-speed broadband can be delivered to anyone, anywhere — rapidly and economically.”

G1 early adopters include large, established telecom service providers such as MTN, BT, and Liberty; new entrants on the broadband scene (not yet public); and a fast-growing number of innovative wireless internet service providers including Wisper Internet, Redzone, and Resound Networks.

Basil Alwan concluded, “Tarana is now fully funded to help close the digital divide with our operator partners worldwide. Our founding engineers set out to make real progress against this intractable challenge, and after years of effort it is gratifying to see such traction in the market!”

http://www.taranawireless.com 

GBI deploys Infinera’s ICE Optical Engines

Gulf Bridge International (GBI), a global cloud, connectivity, and content enabler, deployed Infinera’s Infinite Capacity Engine (ICE) technology across its submarine and terrestrial networks serving the Gulf Cooperation Council (GCC) countries.  The upgrade has enabled GBI to launch new initiatives like its Capacity Protection Program and prepare for major events like the World Cup scheduled to take place in Qatar later this year.

Infinera’s ICE technology on the GX Series Compact Modular Platform has enabled GBI to launch its Capacity Protection Program, which seeks to provide customers with increased protection and optimize both the restoration and repair processes, as well as take a proactive approach to fault prediction and detection. Launched earlier this year, GBI’s program will provide billions of international viewers with a seamless viewing experience of the world’s largest football competition.

“GBI operates significant subsea and terrestrial networks in the Middle East and globally, providing their customers with industry-leading network services,” said Nick Walden, Senior Vice President, Worldwide Sales, Infinera. “By deploying Infinera’s high-capacity coherent optical transmission solutions on their Smart Network, GBI is able to increase network capacity and deliver new services to their customers. We are delighted to support GBI and their customers throughout the Middle East and globally.”

https://www.infinera.com/press-release/gbi-deploys-infinera-ice-optical-engines

GigeNET deploys ADVA FSP 3000 open optical transport

GigeNET has deployed ADVA's FSP 3000 open optical transport technology to meet growing enterprise cloud service demand across the Chicago metropolitan area.

GigeNET’s new DCI infrastructure features full redundancy and optical path protection for enhanced resilience and availability. It’s built on ADVA FSP 3000 open optical transport technology.

“We’ve worked in tight collaboration with GigeNET to deliver a fast and reliable DCI solution that perfectly meets its requirements and ambitions. This new network features enhanced traffic protection, reduces complexity and also provides efficiency savings that support GigeNET’s sustainability drive,” commented John Scherzinger, SVP of sales, Americas at ADVA. “With its upgraded solution, GigeNET can provide its business customers with the peace of mind that comes from fully redundant services. And, as well as always-on availability, our innovation ensures that data takes the fastest networking route to minimize lag and latency.”

https://www.adva.com/en/newsroom/press-releases/20220322-gigenet-meets-soaring-enterprise-cloud-demand-with-advas-dci-solution

Spectro Cloud raises $40M for Kubernetes management platform.

Spectro Cloud, a start-up based in Santa Clara, California, announced $40 million in Series B funding for its Kubernetes (K8s) management platform.

Spectro Cloud’s next-generation Palette Kubernetes management platform delivers full-stack lifecycle management. That includes any combination of new and existing Kubernetes clusters, be they virtualized or bare metal data centers, public clouds or edge locations.

Spectro Cloud announced its Real Metal architecture in October, and its extension to add support for edge locations on March 15. The company’s Palette Edge, Palette Bare Metal, Palette Cloud, Palette Standard and Palette Premium products allow organizations to deploy and manage any type of Kubernetes cluster across any location.

The new funding round was led by New York City-based Stripes, a leading investor in software and consumer products. Other investors participating in the round included Sierra Ventures, Boldstart Ventures, WestWave Capital, T-Mobile Ventures and Alter Venture Partners, all of which also participated in the $20 million Series A funding round that Spectro Cloud closed in July. Additionally, TSG has invested in this Series B round.

“This investment validates Spectro Cloud’s vision and accomplishments toward making Kubernetes more accessible and manageable so organizations can speed up adoption in any environment,” said Tenry Fu, Spectro Cloud co-founder and CEO. “Our approach with Palette is simple: to remove the common pitfalls and trade-offs that organizations face when adopting Kubernetes. Spectro Cloud’s recent Palette Edge and Palette Bare Metal announcements demonstrate our ability to extend a unique approach to every environment under management — even the most challenging ones. The fresh capital will enable us to develop more features, and expand our go-to-market plan internationally to address customer demand.”

https://www.spectrocloud.com 



T-Mobile Ventures invests in SignalWire for APIs

T-Mobile Ventures has made an equity investment in SignalWire, a start-up based in Palo Alto, California, 

SignalWire offers APIs and SDKs that enable developers to reach <50ms latency in large-scale video and voice applications. Its APIs align  with the low latency, high speeds, and broad national coverage of T-Mobile's 5G network.

 T-Mobile, the first U.S. operator to invest in SignalWire, joins Deutsche Telekom and others collaborating with SignalWire to equip developers worldwide with voice, video and messaging APIs to build modern communications applications.

"T-Mobile's leadership in 5G along with its vision and bold set of strategies make them an ideal partner," said Anthony Minessale, Co-Founder and CEO of SignalWire. "We have a shared vision to execute the digital transformation of telecommunications and that starts with arming developers with powerful tools."

"T-Mobile is thrilled to be the first US wireless provider to invest in SignalWire. As we shared in our 5G Forward announcement, we envision a future where everything that can be connected, will be. And that requires massive innovation," said Rob Roy, SVP of Emerging Products at T-Mobile. "SignalWire's solutions are all about enabling developers to build better and faster. This investment is another step in unlocking innovation on our industry-leading 5G network."

http://www.signalwire.com  

  • In June 2021, SignalWire had secured $30 Million in the first closing of a Series B round.

5G forecast: 1.3 Billion by Year-End 2022

According to the most recent data from Omdia, the world added 303 million 5G connections year over year, representing a 139% increase from 218 million to 521 million, and sequentially 19% growth quarter over quarter from 437 million in Q3 2021 to 521 million in Q4 2021. According to Omdia’s forecast, 5G is expected to double to 1.3 billion connections in 2022, nearly double again in 2023 to 2 billion connections, and reach 4.8 billion by the end of 2026. Of that, 516 million is expected to come from North America and 301 million from Latin America and the Caribbean.

Chris Pearson, President of 5G Americas said, “We are now out of the opening stages of this generation of wireless cellular technology, as 5G is rapidly getting into the hands of consumers and businesses, who are finding innovative new ways to use mobile connectivity.”

Omdia’s Principal Analyst Kristin Paulin said, “During 2021 5G coverage really expanded, making 5G an option for a lot more people. The major operators in the US have more than 70% coverage. In Canada, 5G really expanded for the three major operators during 2021, with each ending the year with 70% coverage. This is compared to between 24% and 45% coverage, depending on the operator, at the same time last year.”

By region, North America had a total of 72 million 5G connections by the end of 2021, which is an addition of 54 million 5G connections for the year – a gain of 292 percent year over year. Additionally, the region had 514 million LTE connections by the end 2021, which represents near full market maturity.

Overall, the number of 5G commercial networks has reached 216, according to data from TeleGeography and 5G Americas. That number is expected to reach 330 by the end of 2022 and 352 by the end of 2024, representing strong 5G network investment growth in many regions throughout the world.

https://www.5gamericas.org/5g-forecast-1-3-billion-by-year-end-2022/


Dell'Oro: SASE hit 37% growth in 2021

The convergence of WAN enterprise networking and security fueled strong growth in SASE related networking and security technologies, according to a new report from Dell'Oro Group, with the overall growth rate hitting 37%..

“The pandemic made remote work and cloud-based applications necessary, and by doing so, accelerated the obsolescence of the classic hub-and-spoke networking model,” said Mauricio Sanchez, Research Director, Network Security, and SASE & SD-WAN at Dell’Oro Group. “Rather than thinking of networking and security as separate problems to solve, they are now being thought of as a continuum and driving together cloud-friendly networking and security technologies into SASE,” added Sanchez.

Additional highlights from the 4Q 2021 Network Security Quarterly Report:

  • Over 30 vendors compete in the SASE market, but the top ten took in nearly 80 percent of the revenue.
  • Full-year 2021 market revenue for SASE networking consisting of SD-WAN technologies exceeded $2 billion.
  • Full-year 2021 market revenue for SASE security consisting of Secure Web Gateway, Cloud Access Security Broker, Zero Trust Network Architecture technologies surpassed $2 billion.
  • Unified SASE–a subset of the total SASE market consisting of solutions that implement SASE networking and security as an integrated platform–grew nearly 50 percent year-over-year for full-year 2021.
  • Firewall revenue rebounded in 2021 by rising 13 percent and set a new record exceeding $10 billion for the first time.

https://www.delloro.com/news/sase-alters-market-landscape-with-37-percent-growth-in-2021-according-to-delloro-group/

Motorola Solutions acquires TETRA Ireland

Motorola Solutions has acquired TETRA Ireland Communications Limited, the provider of Ireland’s National Digital Radio Service (NDRS). Financial terms were not disclosed.

TETRA Ireland delivers mission-critical voice and data land mobile radio (LMR) communications to thousands of first responders and frontline workers from national security and enforcement agencies, health and emergency services, state utilities and volunteer organizations across the country. The network is based on the latest TETRA digital radio technology from Motorola Solutions.

“Delivering, managing and operating critical LMR networks is foundational to what we do,” said Greg Brown, Chairman & CEO, Motorola Solutions. “As a trusted technology partner in Ireland for over 40 years, we look forward to continuing to provide the interoperable mission-critical communications that Ireland's emergency and public service agencies rely on every day.”



Tuesday, March 22, 2022

NVIDIA unveils 51.2T Spectrum-4 switch ASIC

NVIDIA unveiled its next generation, 400 Gbps Spectrum-4 Ethernet switch with adaptive routing and enhanced congestion control mechanisms.

The NVIDIA Spectrum-4 uses a custom ASIC with a 51.2 terabits per second switching capacity, supporting up 128 ports of 400G or 64 ports of 800G, and based on 100 Gbps PAM4 SerDes technology. It integrates a 12.8 Tb/s crypto engine wtih support for MACsec and VXLANsec. It also support secure boot as default via hardware root of trust .

NVIDIA Spectrum-4 features a fully-shared and monolithic packet buffer that’s dynamically available to all ports. This provides  microburst absorption with true, port-to-port, cut-through latency. Spectrum-4 also supports programmability of the pipeline and packet modifier/parser without impact tolatency or packet rate, enabling flexibility without compromising performance.

NVIDIA also notes that its Spectrum-4 switch enables accelerated RoCE-based data transport and load balancing through adaptive routing (including adaptive routing notifications for dynamic flow rebalance), as well as high precision congestion control (HPCC) facilitated through in-band network telemetry.

The chip is based on a 4nm process and is expected to sample later this year.

"A new era of massive-scale cloud technologies, such as Omniverse, requires a transformation of data center architecture,” said Kevin Deierling, vice president of Networking at NVIDIA. “The Spectrum-4 platform’s extreme performance and robust security will equip data centers to power breakthrough discoveries that push the boundaries of what’s possible for the benefit of society.”

https://nvdam.widen.net/s/lxhqbqlbqh/ethernet-switches-product-brief-gtc22-spring-spectrum-4-2169045-r3

NVIDIA opens its NVLink die-to-die and chip-to-chip

NVIDIA will expand the use of its NVLink chip-to-chip and die-to-die interconnect technology in its GPUs, CPUs, DPUs, NICs and SOCs. The company also plans to open the technolgy to other for for custom chip and chiplet integrations. 

NVIDIA NVLink-C2C is built on top of NVIDIA’s world-class SERDES and LINK design technology, and it is extensible from PCB-level integrations and multichip modules to silicon interposer and wafer-level connections, delivering extremely high bandwidth while optimizing for energy and die area efficiency.

In addition to NVLink-C2C, NVIDIA will also support the developing Universal Chiplet Interconnect Express (UCIe) standard. Custom silicon integration with NVIDIA chips can either use the UCIe standard or NVLink-C2C, which is optimized for lower latency, higher bandwidth and greater power efficiency.

Some of NVLink-C2C’s key features include:

● High Bandwidth – supporting high-bandwidth coherent data transfers between processors and accelerators

● Low Latency – supporting atomics between processors and accelerators to perform fast synchronization and high-frequency updates to shared data

● Low Power and High Density – using advanced packaging, it is 25x more energy efficient and 90x more area-efficient than PCIe Gen 5 on NVIDIA chips

● Industry-Standard Support – works with Arm’s AMBA CHI or CXL industry-standard protocols for interoperability between devices

“Chiplets and heterogeneous computing are necessary to counter the slowing of Moore’s law,” said Ian Buck, vice president of Hyperscale Computing at NVIDIA. “We’ve used our world-class expertise in high-speed interconnects to build uniform, open technology that will help our GPUs, DPUs, NICs, CPUs and SoCs create a new class of integrated products 

NVIDIA estimates that its NVLink-C2C interconnect could deliver up to 25x more energy efficiency and be 90x more area-efficient than PCIe Gen 5 on NVIDIA chips and enable coherent interconnect bandwidth of 900 Gbps or higher.

“As the future of CPU design is increasingly accelerated and multichip, it is critical to support chiplet-based SoCs across the ecosystem,” said Chris Bergey, senior vice president and general manager of the Infrastructure Line of Business at Arm. “Arm is supporting a broad set of connectivity standards and designing our AMBA CHI protocol to support these future technologies, including collaborating with NVIDIA on NVLink-C2C to address use cases like coherent connectivity between CPUs, GPUs and DPUs.”

https://www.nvidia-press.com/f/preview/27653