Tuesday, March 28, 2023

Enfabrica unveils multi-terabit Accelerated Compute Fabric chips

Enfabrica, a start-up based in Mountain View, California, emerged from stealth to unveil a new class of Accelerated Compute Fabric (ACF) chips optimized for Artificial Intelligence (AI) and accelerated computing workloads.

Enfabrica's Accelerated Compute Fabric (ACF) devices are designed to deliver multi-terabit-per-second data movement between GPUs, CPUs, accelerators, memory and networking devices. The company envisions composable AI fabrics of compute, memory and network resources that scales from a single system to tens of thousands of nodes. The design provides uncontended access to >50X DRAM expansion over existing GPU networks via ComputeExpressLink (CXL) bridging.

Enfabrica’s said its flagship ACF switch silicon enables customers to cut their cost of GPU compute by an estimated 50 percent for Large Language Model (LLM) inferencing and 75 percent for deep learning recommendation model (DLRM) inferencing at the same performance point.

Enfabrica’s first Accelerated Compute Fabric Switch (ACF-S) chip offers multi-port 800 Gigabit Ethernet networking and high-radix PCIe Gen5 and CXL 2.0+ interfaces. Without changing physical interfaces, protocols or software layers above device drivers, the ACF-S delivers multi-Terabit switching and bridging between heterogeneous compute and memory resources in a single silicon die, while dramatically reducing the number of devices, I/O latency hops, and device power in today’s AI clusters consumed by top-of-rack network switches, RDMA-over-Ethernet NICs, Infiniband HCAs, PCIe/CXL switches, and CPU-attached DRAM.

Notably, by incorporating unique CXL memory bridging capabilities, Enfabrica’s ACF-S is the first data center silicon product in the industry to deliver headless memory scaling to any accelerator, enabling a single GPU rack to have direct, low-latency, uncontended access to local CXL.mem DDR5 DRAM at more than 50 times greater memory capacity versus GPU-native High-Bandwidth Memory (HBM).

“Generative AI is rapidly transforming the nature and volume of compute traffic in data centers,” said Rochan Sankar, CEO and co-founder of Enfabrica. “AI problem sizes and user serving scale will continue to grow by orders of magnitude every couple of years. The problem is that current server I/O and networking solutions have serious bottlenecks that will cause them to either buckle under the scale of demand, or vastly underutilize the costly compute resources deployed to meet it. We believe we have cracked the code on a I/O fabric design that will scale high-performance, low-latency AI workloads at far superior TCO than anything out there, and make growing AI infrastructure composable, sustainable and democratized. We are excited to work with our partners and target customers to deliver our Accelerated Compute Fabric products into this fast evolving market.”


  • Enfabrica began in 2020, funded by Sutter Hill Ventures, and led by Silicon Valley veterans Rochan Sankar, Shrijeet Mukherjee, and key engineers who built industry-leading silicon and software stacks at Broadcom, Google, Cisco, AWS and Intel.

Graphiant lands $62 million for its network edge service

Graphiant, a start-up based in San Jose, California, announced a $62 million Series B funding round for its network edge service that promises to shake up the established framework of VPNs, including MPLS and SDWAN.

Graphiant provides a next-gen network edge that combines the performance and security of MPLS with the agility of "as-a-Service" delivery. The company says it can deliver SLA-class performance at up to 70% less cost. Graphiant addresses three important use cases. First, enterprises can use Graphiant to connect enterprise resources. Second, they can connect to multiple public clouds. And third, enterprises can connect to other businesses, such as partners or customers.

The round brings the company's total funding to $96 million.

"The reaction to Graphiant's network edge has been swift and enthusiastic," said Khalid Raza, founder of Graphiant. "It has been much faster than what we saw with MPLS at Cisco or SD-WAN with Viptela. Graphiant has hit a nerve.”


Video: What's next for enterprise connectivity? (after MPLS and SDWAN)

What's next for enterprise connectivity?Graphiant, a start-up based in Silicon Valley, has just unveiled a new network edge service that promises to shake up the established framework of VPNs, including MPLS and SDWAN.In this video, we hear from Khalid Raza, founder and CEO of Graphiant, and who is widely regarded as a visionary in routing protocols and large-scale network architectures. Raza previously was founder & CTO of Viptela, an SDWAN...

NTT develops Beluganos, a network OS for carrier white boxes

NTT announced "Beluganos", a network OS for carrier white boxes.

The technology was developed with IP Infusion and in alignment with NTT's IOWN vision.

Beluganos includes various advanced functions, such as routing and switching capabilities that can replace legacy equipment with a choice of open networking hardware and software solutions, in addition to NTT’s commercial network operation expertise.

Following verification of the network's high reliability, NTT Communications decided to introduce Beluganos into their commercial network infrastructure in combination with white box equipment to support their network services. Operation of the service will commence on March 31, 2023. 

From then on,  NTT Advanced Technology Corporation will provide sales and support for the Beluganos service.

IP Infusion says it will also work to scale the IOWN concept around the world through global sales of Beluganos.

“We are incredibly proud to support NTT in the development, go-to-market and distribution of Beluganos,” said Atsushi Ogata, CEO and President of IP Infusion. “This project demonstrates our ongoing commitment to fostering broad access and high quality software solutions that further innovation and lower total cost of ownership.”


Microsoft brings OpenAI to cybersecurity

Microsoft launched private previews of a next generation cybersecurity tool with access to OpenAI models.

The company says its aim is to amplify the capabilities of security teams by summarizing and making sense of threat intelligence, helping defenders see through the noise of web traffic and identify malicious activity.

Microsoft Security Copilot correlates input from the company’s vast threat intelligence footprint with expertise from security professionals and other tools, including Microsoft Security products, such as Microsoft Sentinel and Microsoft Defender.

“Today the odds remain stacked against cybersecurity professionals. Too often, they fight an asymmetric battle against relentless and sophisticated attackers,” said Vasu Jakkal, corporate vice president, Microsoft Security. “With Security Copilot, we are shifting the balance of power into our favor. Security Copilot is the first and only generative AI security product enabling defenders to move at the speed and scale of AI.”

“Advancing the state of security requires both people and technology — human ingenuity paired with the most advanced tools that help apply human expertise at speed and scale,” said Charlie Bell, executive vice president, Microsoft Security. “With Security Copilot we are building a future where every defender is empowered with the tools and technologies necessary to make the world a safer place.”


DOCOMO deploys Nokia’s 7750 SR-14s core routers

NTT DOCOMO has deployed Nokia’s 7750 SR-14s core routers, powered by FP routing silicon, on its nationwide IP core backbone to enable transport network slicing as it rolls out new 5G mobile services. The enhancement will allow granular SLAs, network scale, capacity and agility, along with increased power and resource efficiency.

Nokia's FP5 line cards provide 800GE capability, increase capacity by more than three times with a 75 percent power savings over previous generations, and simplify network evolution with concurrent line rate FP4 and FP5 line card operation in the same system. To deliver the network slicing for its mobile services, DOCOMO will leverage the comprehensive segment routing capabilities of Nokia’s proven Service Router Operating System (SR OS) for traffic engineered network slices for the granular SLAs required by DOCOMO’s customers.

Nokia’s NSP complements DOCOMO’s segment routing solution with a Path Computation Engine (PCE) that leverages real-time telemetry to automatically optimize the IP network and improve SLA adherence. NSP will also enable DOCOMO to automate the creation, assurance, and optimization of IETF standards-based network slices in the transport domain.

Fumitaka Murayama, General Manager of Core Network Engineering Department, NTT DOCOMO, INC., said: “DOCOMO is committed to implementing slicing in 5G networks to provide networks that can respond flexibly and quickly to the diverse needs of our customers. In the transport network, a key component of this, it was essential to leverage Nokia's IP core solutions for its high performance, functionality, flexibility and quality in meeting performance and slicing requirements. In collaboration with Nokia, DOCOMO hopes to create new value for our customers and help solve social issues through new technologies such as network slicing.”

John Lancaster-Lennox, Head of Market Unit Japan at Nokia, said: “We’re pleased to provide our next generation IP transport solution to DOCOMO. Nokia’s 7750 SR-s platform and NSP solution will enable DOCOMO to balance network scale, capacity and sustainability that can evolve with its changing needs for years to come.”

Namibia's Paratus selects Infinera’s ICE6 800G for Equiano

Paratus, a network operator serving seven African countries – Angola, Botswana, DRC, Mozambique, Namibia, South Africa, and Zambia - has selected  Infinera’s ICE6 800G solution for the Equiano Cable System.

Schalk Erasmus, CEO of the Paratus Group, stresses: “It is imperative to select top technology partners, as is the case in the partnership with Infinera, both for the modernization of the submarine network infrastructure, and for the quality of network provided by the Group to its customers and surrounding communities.” 

“Paratus have worked with Infinera for a number of years on other projects, and this partnership allows Paratus to provide connectivity that is 20 times faster in capacity, modern and flexible to ensure economies of scale, and offers efficiency in the services that the Paratus Group provides throughout its seven operating countries.”


Google announces Equiano cable from Portugal to South Africa

Google unveiled plans for new private subsea cable connecting Lisbon, Portugal and Cape Town, South Africa with branching units along the west coast of Africa.

The Equiano subsea cable, which is named for Olaudah Equiano, a Nigerian-born writer and abolitionist who was enslaved as a boy, will include a branching unit to Nigeria.

Google plans to use state-of-the-art space-division multiplexing (SDM) technology to achieve approximately 20 times more network capacity than the last cable built to serve this region.

The cable will be the first to incorporate optical switching at the fiber-pair level, rather than the traditional approach of wavelength-level switching. Google says this design will greatly simplify the allocation of cable capacity.

Equiano is fully funded by Google. Alcatel Submarine Networks was awarded a contract in Q4 2018.