Tuesday, December 19, 2023

Lightmatter pulls in additional $155M for photonic interconnects / processors

Lightmatter, a start-up based in Boston that is developing photonic interconnect and compute products, annouced $155 million in a Series C-2 funding round. The company has raised over $420 million to date and is now valued at over $1.2 billion. The latest funding round was led by GV (Google Ventures) and Viking Global Investors, with participation from others. 

Lightmatter said its full stack of photonics-enabled hardware and software solutions will reduce power consumption and increase performance.

“Lightmatter is positioned to be a key driver in powering the next generation of computing systems that will further enable AI innovation. Through photonic technologies, Lightmatter is ensuring the steady progress in computing performance continues, despite growing power consumption challenges and slowing progress with transistor scaling,” said Lightmatter co-founder and CEO, Nick Harris. “We are thankful for the support of our investors, each of whom share our vision of Lightmatter playing an integral role in the future of computing. By increasing speed, lowering cost, and reducing environmental impact, our technologies can continue to push the limits of what’s possible, fueling greater AI adoption and innovation.”

Some additional notes:

  • Since May 2023, Lightmatter has grown its headcount more than 50% to meet client demand and product milestones. 
  • Lightmatter plans to open a Toronto office in 2024. 
  • Danner Stodolsky has been appointed Vice President of Data Center Architecture.  Danner was previously Senior Vice President of Cloud at SambaNova and a Vice President of Engineering at Google.
  • Colin Sturt has been appointed General Counsel. He brings over 20 years of semiconductor industry experience to Lightmatter and was most recently Senior Vice President and General Counsel at Dialog Semiconductor.


OpenZR+ MSA tests multi-vendor interoperability

 The OpenZR+ Multi-Source Agreement (MSA) Group announced the results of two successful interoperability test events at LightRiver and featuring coherent optical transceivers from Cisco, Coherent, Fujitsu, Juniper Networks, and Lumentum operating on Cisco and Juniper routers. 

The testing demonstrated compatibility to the OpenZR+ specification and interoperability between optical transceiver modules from different vendors in two different routers by transporting 400 Gbps traffic over various optical links. 


  • Phase 1 (August 2023): Verification that optical transceiver QSFP-DD-DCO modules comply to the OpenZR+ multi-source agreement in router platforms through the OIF CMIS compliant interface in loopback mode. Also verify that optical transceiver QSFP-DD-DCO modules from different vendors interoperate over a Single-Span link with 75 km of fiber.
  • Phase 2 (September 2023): Multi-vendor Interoperability and performance testing over multi-span optical line system (OLS).

The OpenZR+ MSA Group also published a whitepaper that notes:

  • 400G OpenZR+ optical transceiver modules from multiple vendors operated in different routers
  • All 400G OpenZR+ optical transceiver modules transported traffic with no post-FEC bit errors at receiver OSNR of 24 dB or less
  • 400G OpenZR+ optical transceiver modules from five different vendors interoperated with each other
  • 400G traffic was transported over a typical DCI use case with 75 km of fiber and a typical metro use case with multi-spans over 430 km of fiber.

"The results of this multi-vendor interoperability testing validate the readiness of network operators to take advantage of 400G pluggables in order to cost-effectively scale their networks to meet growing bandwidth demands," said Tom Williams, Co-chair of the OpenZR+ MSA Group.

"Not only does interoperability enable a more robust supply chain in the industry by allowing network operators to mix and match equipment and services from different providers, but it can also accelerate the adoption of new technology," said Atul Srivastava, Co-chair of the OpenZR+ MSA Group.

The whitepaper can be downloaded from the OpenZR+ MSA Group website.


  • in September 2023, the OpenZR+ Multi-Source Agreement (MSA) Group published the OpenZR+ Rev 3.0 specification which defines a higher performing 400G 8QAM mode as well as a higher transmit power mode.The new specification is designed to further expand the application space for a coherent solution in small form factor pluggable modules, such as form factors defined by the QSFP-DD MSA and OSFP MSA.

Data Center Architecture Spanning Edge to Cloud with AI & 5G

Check out AI and Cloud Data Center Networking videos here: https://ngi.fyi/aidcnet23yt

AI is making a transformative impact on data center architecture. Yet developers know that applications must flow seamlessly from core to edge. How can these forces be resolved? Sanjay Kumar, VP of Products and Marketing from Arrcus, explains:

  • How Arrcus' solutions optimize and maximize the utilization of GPUs in data centers, minimizing latency and loss in running AI/ML workloads. They offer two architectures: traditional IP CLOS leaf and spine, and a distributed architecture for massive scale,
  • The role of FlexMCN, a network fabric overlay, in stitching together workloads from the edge through the core and into multicloud or public cloud environments. This robust, secure, and flexible solution allows network operators to run workloads wherever they reside and deliver applications wherever they're required.
  • The introduction of an innovative technology, SRv6 mobile user plane, developed in collaboration with SoftBank Mobile. This technology simplifies the delivery of network slicing, lowers the total cost of ownership, and helps network operators monetize their infrastructure.

Additionally, Arrcus' ArcIQ solution provides end-to-end network visibility, streaming real-time telemetry from the underlying infrastructure to give operators a comprehensive view of their network.

Join Sanjay as he takes you through Arrcus' vision for a distributed AI and 5G world, and how their solutions are helping network operators deliver on this vision.

Want to be involved our video series? Contact info@nextgeninfra.io

Arrcus joins the Ultra Ethernet Consortium

Arrcus has joined the Ultra Ethernet Consortium (UEC), a Linux Foundation initiative dedicated to developing next-generation Ethernet technologies for AI and HPC applications.

The UEC is a collaborative effort among industry leaders to address the growing demands of AI and HPC workloads on data center networks. These workloads require high performance, low latency, and scalable interconnects to efficiently move massive datasets and rapidly process complex algorithms.

Arrcus said it is joining the UEC to contribute its expertise in advanced networking technologies, including its Leaf-Spine Fabric Architecture and its support for high-speed Ethernet interfaces, such as 400GE and 800GE. The company is also committed to working with other UEC members to develop and standardize new Ethernet technologies that meet the specific needs of AI and HPC applications.

“We are thrilled to join the UEC and contribute to the development of next-generation Ethernet technologies for AI and HPC,” said Shekar Ayyar, chairman and CEO at Arrcus. “The UEC’s mission to create an open, interoperable, and high-performance Ethernet ecosystem aligns perfectly with our own goals at Arrcus, and we are confident that our participation will help to accelerate innovation in this critical area.”

Arrcus also brings its recently announced ACE-AI solution for distributed AI workloads. This networking solution based on ArcOS, delivers a modern, unified network fabric for optimizing distributed GPU and other compute resources with maximum performance for AI/ML workloads, and positions Arrcus as a significant contributor to the consortium's mission. This solution encompasses next-generation datacenter networking architectures, as well as innovations in hybrid/multi-cloud connectivity and 5G transport to seamless connect GPU resources where ever they may reside.


Comcast confirms massive breach impacting nearly 36m customers

Comcast confirmed a data security breaching involving nearly 36 million of its Xfinity broadband customers across the U.S.

Comcast blamed the cyber intrusion on a zero-day vulnerability in Citrix software. Comcast claims to have patched the software promptly after Citrix issued in mitigation

guidance on October 23, 2023, however, it discovered unauthorized access to its systems between October 16 and October 19, 2023.

It is believed that compromised data included usernames, hashed passwords, contact information, last four digits of social security numbers, dates of birth and/or secret questions and answers.

Comcast's statement is posted here:


Canada's Western James Bay Telecom Network selects Infinera

Western James Bay Telecom Network (WJBTN) selected Infinera to expand connectivity services to rural communities in northern Ontario, Canada. 

The deployment will use the Infinera XTM Series to deliver flexible, reliable, and cost-effective high-bandwidth solutions to its residential, business, and government customers. FONEX Data Systems is Infinera's integration partner.

WJBTN has been bringing telecommunications services to rural communities in Canada from Moosonee along the western James Bay coast through Fort Albany, Kashechewan, and Attawapiskat since 2010. With the selection of Infinera, WJBTN will be able to cost-effectively deliver critical next-generation high-capacity network services to support the region’s economic development and access to education and telehealth.

“FONEX’s unwavering support, coupled with Infinera’s excellent XTM solution, has greatly improved our network design, enabling us to substantially boost bandwidth, expand our service network footprint, and significantly enhance service quality and bandwidth speeds for our valued customers,“ said David Coté, General Manager, Western James Bay Telecom Network.

“We are pleased to support Western James Bay Telecom Network’s upgrade, helping them expand their broadband networking infrastructure with Infinera’s leading technology to deliver robust connectivity services to rural Ontario communities,” said Nick Walden, Senior Vice President, Worldwide Sales, Infinera.

University of Stuttgart picks HPE for supercomputers

The High-Performance Computing Center of the University of Stuttgart selected Hewlett Packard Enterprise to build two new supercomputers. The total combined cost of the project is €115 million and funding will be provided through the Gauss Centre for Supercomputing (GCS), the alliance of Germany's three national supercomputing centers. Half of this funding will be provided by the German Federal Ministry of Education and Research (BMBF), and the second half by the State of Baden-Württemberg's Ministry of Science, Research, and Arts.

In the first stage, a transitional supercomputer, called Hunter, will begin operation in 2025. This will be followed in 2027 with the installation of Herder, an exascale system that will provide a significant expansion of Germany’s high-performance computing (HPC) capabilities. 

Hunter will raise HLRS’s peak performance to 39 petaFLOPS (39*1015 floating point operations per second), an increase from the 26 petaFLOPS possible with its current supercomputer, Hawk. It will use the AMD Instinct MI300A accelerated processing unit (APU), which combines CPU and GPU processors and high-bandwidth memory into a single package.