Mellanox drives NVIDIA into the Data Center

Above: Eyal Waldman, CEO and President of Mellanox (left) and Jensen Huang, founder and CEO of NVIDIA

The $6.9 billion merger between Mellanox and NVIDIA which was completed in April 2020, proved to be a successful move: Mellanox is an expert in high-speed Ethernet and InfiniBand networking solutions for Data Centers and High Performance Computers. NVIDIA invented the GPU back in 1999, and transformed itself from a provider of graphical boards into a powerhouse in High Performance computing and Artificial Intelligence, based on its expertise in GPUs.

This newly wedded couple produced an excellent financial quarter in Q2 2020: Record revenue of $3.87 billion, up 50% from a year earlier, and record Data Center revenue of $1.75 billion, up 167% compared to Q2 2019. Mellanox contributed approximately 14% of company revenue and just over 30% of data center revenue. In fact, Mellanox’ own sales grew 26% under the umbrella of NVIDIA to a record of $541 million.

“That’s why we bought Mellanox”

Mellanox also played an important role in NVIDIA’s strategy to break into the data centers market. “Mellanox grew sharply, driven by the need for high-speed networking in cloud data centers to scale-out AI services,” said Jensen Huang, President and Chief Executive Officer of NVIDIA. “We recognize the importance of high-speed networking and low-latency networking, and that’s why we bought Mellanox.”

Colette Kress, Chief Financial Officer of NVIDIA said during a conference call earlier this week, that NVIDIA and Mellanox powers two-thirds of the top 500 Super Computer systems in the world,  compared with just less than a half in total two years ago. “And just this morning, Microsoft Azure announced the availability of massively scalable AI clusters, which are based on the A100 and interconnected with 200-gigabyte-per-second Mellanox InfiniBand networking.”

Mellanox' ConnectX-6 Dx Dual 100GbE / Single 200GbE SmartNIC for Advanced Cloud
Mellanox’ ConnectX-6 Dx Dual 100GbE / Single 200GbE SmartNIC for Advanced Cloud

The overall strategy is well defined: “The combination of NVIDIA accelerated computing, Mellanox networking, and Cumulus software (Acquired in May 2020) – enables data centers that are accelerated, disaggregated, and software-defined – to meet the exponential growth in AI, cloud, and high-performance computing.”

Growth to continue in Q3 2020

Jensen sees more sales in the near future. “Despite the pandemic’s impact on our professional visualization and automotive platforms, we are well positioned to grow, as gaming, AI, cloud computing and autonomous machines drive the next industrial revolution,” he said, and expects Q3 2020 revenue to be $4.40 billion, plus or minus 2%.

Why? Because “Two components, two types of technologies are really important to the future of cloud. One of them is acceleration, and our GPU is ideal for it. And then the other one is high-speed networking. That transition is called east-west traffic. And the most important thing you could possibly do for yourself is to buy really high-speed, low-latency networking. And that’s what Mellanox is fantastic at.”

Nvidia and Mellanox built a Supercomputer in just a Month

Photo above: Mellanox’ AI platform protects supercomputers from from hacking and inappropriate use

In a first joint announcement by Nvidia and Mellanox, the two companies announced a reference design for the rapid building of supercomputers, and a new cyber protection platform for supercomputers. Mellanox has expanded its offering of Unified Fabric Manager (UFM) products, adding to it a new appliance called UFM Cyber-AI Platform.

It provides cyber protection to supercomputers and big data centers, using an artificial intelligence software that studies the behavior characteristics of the computing systems, to identify malfunctions and detects abnormal activity that implies on hacking and unauthorized activity.

Originally, UFM technology was developed a decade ago by Mellanox in order to manage InfiniBand-based communications systems by providing network telemetry data, monitoring the activity of all the related devices, and managing the software updates across the network’s components.

The new solution comes both as a software package or as a complete appliance based on Nvidia’s dedicated server. It is focused on characterizing computer operation and identifying unusual activity. According to Nvidia and Mellanox, the system significantly reduces the data center’s downtime, whose damages are estimated to reach $300,000 per hour.

Supercomputers are open and unprotected platforms

According to Mellanox’s VP of Marketing, Gil Shainer, the integration of Mellanox’s InfiniBand with Nvidia’s GPU changes the rules of the game in the supercomputer market, bringing to it unprecedented cyber security and preventative maintenance capabilities. Shainer: “Supercomputers are managed differently from organizational computer centers. Usually it is an open platform that need to provide easy access to many researchers around the world.”

To illustrate the dilemma he recalled an event that took place several years ago at an American university. “The administrator of the computers center told me how they caught a student using a computer for crypto mining. The suspicion emerged when they found out that the computer’s power consumption was not declining during the annual vacation, a period of time in which the computer usually is not active. Our solution allows you to detect such a situation right away – and not have to wait for your computer’s power bill.”

Reference Design for the Rapid Construction of Supercomputer

Alongside the joint announcement, Nvidia unveiled a new supercomputer called Selene (photo above), which is considered the strongest industrial supercomputer in the United States, with peak performance of 27.5 petaflops. The computer is based on the new A100-model GPU processors announced this week, and was built for internal research conducted in Nvidia. During a press briefing last week, Shainer revealed that the new computer was built in just one month, a record-breaking time for the construction of a supercomputer.

Shainer: “The ability to build a supercomputer in a month is based on expertise in communication and expertise in processors. We have developed a reference design that allows anyone to build a supercomputer, based on ready made blocks of Nvidia’s processors and Mellanox’s communication. Because the processors are fully compatible with the communications cards, the computer can be set up in no time. In fact, we have jointly developed a reference design that allows for the construction of computers of any size – not just supercomputers.”

Mellanox to Acquire Titan IC from Northern Ireland

Mellanox Bluefield Chip

Mellanox Technologies has reached a definitive agreement to acquire Titan IC, to strengthen its network intelligence capabilities delivered through the company’s ConnectX and BlueField families of SmartNIC and I/O Processing Unit (IPU) solutions. Titan IC’s 30 employees team in Belfast, Northern Ireland will become the center of advanced network intelligence research and development for Mellanox. Finencial terms were undisclosed. Israeli press speculations mentioned $15-25 million value of Titam IC.

Titan IC provides high-speed complex pattern matching and real-time Internet traffic inspection for advanced cyber security and data analytics applications. Its RXP hardware network intelligence acceleration engine, offloads regular expressions (RegEx) string and malware matching, freeing up CPU cores and delivering faster throughput and increased efficiency. The RXP engine’s unique ability to simultaneously handle high throughput, rule depth, and complexity is complementing Mellanox’ BlueField IPU based solutions.

“Network Intelligence is an important technology when combined with our industry-leading networking portfolio of switches, SmartNICs, and IPUs,” said Eyal Waldman, president and CEO Mellanox Technologies. “ With this acquisition, our M&A investments total more than 1.2 billion dollars to date and more than 53 million dollars invested in startups to further our intelligent networking strategy.”

Noel McKenna, CEO of Titan IC, said that the company have worked with Mellanox for many years to integrate its RXP regular expression processor into BlueField I/O Processing Units (IPUs). “Now as part of Mellanox, we will be able to achieve new capabilities for cyber intelligence, intrusion detection and protection, and advanced data analytics applications.”


Mellanox Has Gained Over a Million Ethernet Ports

 Mellanox Technologies from Yokneam, Israel, said that the integration of its Ethernet switches into  Check Point Software’s “Maestro” security platform, announced in January 2019, has propelled Ethernet switch shipments past the one million port milestone. Check Point has selected Mellanox Spectrum-based Ethernet switches for its next generation Maestro platform – the industry’s first truly hyperscale network security solution.

Check Point Maestro is a revolutionary new hardware-software security suite based on a new architecture to enable cyber defense against “Generation V” threats.  Its HyperSync patented technology allows customers to enjoy full redundancy within a system, by utilizing all hardware resources. Maestro leverages the Mellanox 10, 25 and 100Gbps Spectrum based Ethernet switches.

The Spectrum-based Open Ethernet platform enables Check Point to allow Gaia, its operating system for security applications, to run on the switch. Gaia provides the feature set and interfaces that are familiar to security administrators worldwide, and delivers secure updates via IPSO and Secure Platform.

Mellanox’ Spectrum family of Ethernet switches are purpose built for Software Defined Datacenters. The Spectrum product family includes a broad portfolio of Top-of-Rack (TOR) and aggregation switches that range from 16 to 128 physical ports, with Ethernet data rates of 1GbE, 10GbE, 25GbE, 40GbE, 50GbE, 100GbE, 200GbE and 400GbE per port. Utilizing an Open Ethernet architecture allows a choice of operating systems, including Mellanox Onyx, Cumulus Linux, SONiC, Switchdev and others.

Itai Greenberg, VP Product Management at Check Point, said that Check Point’s decision “Came After a careful evaluation of the various offerings. We selected the Mellanox Spectrum family of switches as the best solution to meet our needs and deliver unmatched security platforms to our customers. We look forward to continuing the technology collaboration with Mellanox for product generations to come.”

Mellanox leads the European plaCMOS project

Mellanox Technologies from Yoqneam, Israel, revealed today that it is the coordinator of the plaCMOS is a research project on photonic integration, supported by the Horizon2020 Framework Programme for Research and Innovation of the European Commission. This revolutionary project may shape a new direction for the Silicon Photonics Industry. The goal of the PlaCMOS project is to develop and demonstrate the next generation optical-electronic CMOS platform that will enable transceivers capable of real-time communication with data rates exceeding 200 Gb/s.

PlaCMOS will be based on the latest BiCMOS technology, and co-integrate it with a ferroelectric plasmonic modulator and a SiGe detector technology. By exploiting plasmonics rather than photonics, plaCMOS will not only be able to extend the bandwidth far beyond 100 GHz but also reduce the footprint of Photonics to the micrometer scale.

The wonder of Plasmonics

Plasmonics deals with the interaction between electromagnetic field and free electrons in a metal. Since electrons in thin films metals such as Gold and Silver  can be excited by the electric component of light, it is possible to directly determine the wavelength of light be electric fields created via electric circuitry.

The PlaCMOS project is planning a new chip, that includes all the analog, digital and optics in a single die and will be manufactured in a standard CMOS process. The technology demonstrator will be based on BiCMOS technology, a ferroelectric plasmonic modulator and a SiGe detector. It will be a single channel 200 Gb/s NRZ transmitter and receiver pair, and an ultra-compact 4×50 Gb/s transceiver directly interfacing a multicore fiber. Later, a full switch will be built in Mellanox facility in Israel.

The other participants in the program include Micram Microelectronic GmbH from Germany, The IHP institute from Germany, IBM Research GmbH from Switzerland, ETH Institute from Switzerland, Saarland University (Germany) and The Aristotle University of Thessaloniki (Greece).

For more information: PlaCMOS