Orange Business launches GenAI

Orange Business launched two new trusted generative AI (GenAI) offers covering the infrastructure layer (IaaS) and applications (SaaS). 

SaaS GenAI offer - This fully integrated SaaS solution, which is offered in collaboration with LightOn, is designed for enterprises of all sizes (SMEs and ETIs) who want a ready-to-use service that can adapt to a wide array of use cases (document management, reporting, content generation, etc.). 

• The service was developed in partnership with LightOn, a French startup that provides a complete solution with Large Language Models (LLMs) and intuitive business interfaces, which makes it extremely easy and understandable for average users to use and optimize. 
• Unique to the market, the service can be tailored to different organizations by integrating specific document databases. Customers can benefit from the potential of GenAI with a very low error rate. 
• The solution, entirely managed by Orange Business, combines data confidentiality and security with efficient cost control.

GPU as a Service for inference and training - this IaaS offer is for customers who require complete control over a trusted infrastructure that can host all types of GenAI projects. On its Cloud Avenue platform, Orange Business provides dedicated hardware resources (GPUs) allowing businesses to deploy their own LLMs – with a partner of their choice – for flexibility and sufficient computing power. The solution offers a high degree of customization and addresses two primary use cases:

• Training: fine-tuning and specialization powered by servers equipped with ultra-high-performance GPUs. The infrastructure dedicated for this training is designed to train complex models that require significant computing power.
• Inference: for large-scale GenAI projects. Customers now have access to a dedicated infrastructure for the development and deployment of their models for a large number of users.

The offers are supported by very high bandwidth servers and Graphics Processing Units (GPUs) and are hosted in Orange data centers in France and operated by Orange Business on its trusted Cloud Avenue platform. Orange Business is partnering with with Hewlett Packard Enterprise (HPE),which provides the hardware infrastructure with HPE GreenLake.

https://www.orange-business.com/en/press/orange-business-launches-new-trusted-genai-offers-end-end-french-customers

How is Orange Business leveraging GenAI in its operations and for the benefit of global enterprise customers? Usman Javaid, Chief Products & Marketing Officer from Orange Business, explains:

- The importance of having the right digital infrastructure: Many businesses are not ready for GenAI because they lack the necessary digital infrastructure. This includes having data in the right place, a flexible and adaptable network, and the correct security parameters and data compliance.

- The role of the 3Cs - Cloud, Connectivity, and Cybersecurity: These three elements combine to create a scalable digital infrastructure that is essential for GenAI implementation.

- The potential of GenAI for new use cases: Orange Business is helping customers identify new use cases for Gen, such as integrating Gen capability into contact center solutions for better real-time support, and using GenAI for internal network planning to reduce planning time.

Remember, without the right digital infrastructure, GenAI implementation is impossible. 

https://youtu.be/SO-W_uMXj2M

Read More

Nile implements AI Copilot across its full networking stack

Nile, a start-up offering next-gen enterprise network-as-a-service (NaaS), introduced an AI-driven  architecture for its Nile Access Service. 

Components of the architecture include Nile Service Blocks, Nile Services Cloud, and two categories of AI applications: Nile Copilot and Nile Autopilot. 

The company says its network architecture for enterprise campus and branch locations represents the final frontier for enabling automated operational and service-centric innovation models. Key to this transformation are delivering network connectivity using an as-a-service model, AI for automation and observability, and integrated security. The aim is to go beyond AI-generated report summaries to completely automate the traditionally manual lifecycle of enterprise networks, from design to implementation to ongoing operations.

“Nile was created with a vision to change the way enterprise networks are designed, consumed, and operated. Our mission is to give IT leaders a bridge over the growing innovation gap between their goals and the resources required to achieve them,” said Pankaj Patel, CEO and co-founder, Nile. “The first step in that journey was building the industry’s first service architecture to deliver connectivity with guaranteed performance outcomes. Today, we’re building on that by ushering in a new era powered by AI networking.”

Three core elements enable these comprehensive capabilities: 

Nile Service Blocks - each block represents a collection of physical Wi-Fi sensors, Wi-Fi access points, access switches, or distribution switches. Different sizes of service blocks for different sites are designed automatically and delivered to enable secure connectivity at campus and branch locations given site survey requirements. Instead of requiring manual configuration and separate software release management for different network elements, service blocks are supported with cloud native delivery principles and a microservices-based architecture. Nile Service Blocks vertically integrate more than ten traditionally separate products and are purpose-built to be delivered as a service.

As part of the Nile Service Blocks, the Nile networking hardware is purpose-built for AI, and the software is deeply instrumented across all layers, utilizing deterministic system design and the continuous collection of 1000+ data points across all aspects of the edge infrastructure. Nile Service Blocks integrates zero-trust networking principles and enforces L3-only isolation for all connected devices on hardened hardware elements.

Nile Services Cloud - powered by an integrated data model, Nile Services Cloud enables real-time observability and continuous optimization of Nile Service Blocks by utilizing both model-centric and data-centric AI technologies. It drives the following unique innovations for AI networking:

• Design Pipeline: Automated day -1/0 operations, including survey, design, and installation.
• Digital Twins: Virtual replica of the service blocks, helping simulate network operations.
• Defense Hub: Edge-to-cloud zero-trust security is built into all service blocks.
• Smart Agents: Real-time data collection from the service blocks at high efficiency.
• Cognitive Decisions: Closed loop automation to maintain the best state for the network.

Nile AI Apps -these categories of applications radically improve how system installers, IT admins, end users, and Nile’s own customer success and production engineering teams gain visibility and control across the tenants of the Nile Access Service.

• Nile Copilot applications offer one-touch installation and intent-based provisioning of Nile Service Blocks and enable IT admins and their end users to monitor service outcomes and securely onboard user and IoT devices.
• Nile Autopilot applications offload traditionally manual network operations center (NOC) functions such as software maintenance and eliminate manual workflows for infrastructure performance monitoring and troubleshooting.

https://nilesecure.com/press-releases/nile-introduces-industrys-first-ai-networking-solution-with-performance-guarantees/

Read More

Tech Update: AI Copilots for the Full Network Architecture

 Is network automation ready to take the next big leap? Austin Hawthorne, VP Worldwide Solutions Architecture from Nile, explains:

- How Nile's Close Loop automation, a cloud-native system with a vertically integrated stack, automates everything from Day Zero to ensure predictable outcomes at scale.

- The concept of a 'digital twin' and how it helps in identifying and solving any deviations that could lead to performance issues before they become a problem.

- Real-life scenarios where Nile's automation has detected potential issues, implemented remediation actions, and reduced productivity issues from days and weeks down to seconds and minutes.

https://youtu.be/finZ5jhi6Po

Cloud-based NaaS solutions promise  faster scalability, cost efficiency, security, and streamlined management. We invite you to join our Campus NaaS 2024 Showcase, where we will delve into these questions and more. Reach us at info@nextgeninfra.io

Have a tech update that you want to brief us on? Contact info@nextgeninfra.io!

false

Read More

Cisco and Ribbon demo interoperability at 1.2 Gbps

Cisco and Ribbon Communications confirmed interoperability of their optical transmission systems at 1.2 Tbps.

At the upcoming #OFC24, the companies will demonstrate interoperability between the Ribbon Apollo 9408 and the Cisco NCS 1014, which both incorporate Acacia's CIM 8 140Gbaud transceiver.  The demo will also show conformance to the OpenROADM bookended model for transponder interoperability while uniquely extending this model for multiple vendors.

"We're launching a new era with this demonstration and showcasing a proven interoperable multivendor ecosystem for high performance optical transmission that is a new direction from what has previously been the industry's focus," said Sam Bucci, Chief Operating Officer, Ribbon. "We're proud to collaborate with Cisco on this critical update, giving operators the confidence to access and deliver 1.2 T transmission speeds without resorting to a single closed solution."

"With the growth of AI and other emerging technologies, network operators are challenged with the complexity of managing multi-vendor networks to keep up with the growing demand for faster speeds," said Bill Gartner, SVP/GM Optical Systems and Optics from Cisco. "Together with Ribbon, we are demonstrating that now our customers can take advantage of interoperability at the transmission level before standards have been put forth, enabling them to stay on the cutting-edge without being locked into proprietary solutions."

Read More

Keysight and Credo demo full line rate 1.6 Terabit Ethernet

Keysight Technologies and Credo Semiconductor developed a joint testbed that demonstrates the first IEEE P802.3dj draft specification compliant 1.6 terabit Ethernet measurement system running full line rate layer 2 Ethernet traffic on a real world hardware development platform. 

The demonstration proves that 1.6T Ethernet speed is possible and that its performance can be measured.

Highlights of the Keysight and Credo testbed demonstration include:

• Operational 1.6T Layer 2 engine – The engine operates on an interface of 16 electrical lanes running at 106.25 gigabits per second (Gbps) per lane with Pulse Amplitude Modulation 4 (PAM4) signaling and Forward Error Correction (FEC). This is also called 16 x 100G SERDES.
• Keysight hardware development test platform –With OSFP-XD front panel ports that can send and receive full line rate 1.6T layer 2 bi-directional traffic over a Credo provided 2-meter HiWire Active Electrical Cable (AEC). The 1.6T measurement system measures the performance of the link and the AEC.
• Reliability with FEC – The Layer 2 engine running at full 1.6T line rate provides packet and FEC measurements, plus counters, and other related statistics. FEC enhances data reliability by detecting and correcting errors that occur during data transmission.

Bill Brennan, Chief Executive Officer, Credo, said: “Credo launched the 1.6T HiWire AEC in support of our key customers’ product roadmaps in 2023 and it’s generated a great deal of industry interest. We’re excited to work with Keysight to support the buildout of a robust test and measurement ecosystem for this developing networking standard.”

Ram Periakaruppan, Vice President and General Manager, Network Test & Security Solutions, Keysight, said: “1.6T Ethernet is the next generation of Ethernet that addresses the demand for more network bandwidth and data throughput, driven in part, by ever-increasing AI workloads within the network infrastructure. The 1.6T Ethernet Layer 2 traffic generation and analysis demonstration moves the needle forward to enable development of 1.6T devices and the entire networking equipment ecosystem. Keysight’s Layer 2 and FEC test and measurement systems are also vital to the success of the eventual deployment of 1.6T capable networking devices and equipment.”

Read More

Credo debuts 400G Active Electrical Cables series

Credo Technology is sampling a new family of HiWire Active Electrical Cables (AECs) specifically targeted for 400G AI/ML Backend Network Connections to Top of Rack (TOR) Switches.

The new family of HiWire AECs includes:

• 400G OSFP-RHS (4x112) to 400G QSFP-DD (8x56) for 12.8T TORs
• 400G QSFP112 (4x112) to 400G QSFP-DD (8x56) for 12.8T TORs
• 800G 2xOSFP-RHS (4x112) to 800G OSFP (8x112) for 25.6Tb and 51.2Tb TORs
• 800G 2xQSFP112 (4x112) to 800G OSFP (8x112) for 25.6Tb and 51.2Tb TORs
• 800G 2xOSFP-RHS (4x112) to 800G QSFP-DD (8x112) for 25.6Tb and 51.2Tb TORs
• 800G 2xQSFP112 (4x112) to 800G QSFP-DD (8x112) for 25.6Tb and 51.2Tb TORs
• 800G 2xQSFP-DD (4x112) to 800G QSFP-DD (8x112) for 25.6Tb and 51.2Tb TORs

“400G AI/ML backend networks have proliferated in the past year and have migrated 112G/lane connectivity to Network Interface Cards (NICs) ahead of many customers’ networking plans,” said Don Barnetson, Vice President of Product at Credo. “Credo’s innovative HiWire AECs can help bridge this gap with in-cable speed shifting to enable use of legacy 12.8T TORs and Y cable configurations for 25 and 51Tb TORs.”

Read More

Eutelsat signs with Intelsat's OneWeb LEO constellation

 Eutelsat Group has announced a multimillion-dollar deal with Intelsat for its OneWeb Low Earth Orbit (LEO) Constellation. The agreement, set to commence in mid-2024, is valued at up to $500 million over a span of seven years, with a guaranteed commitment of $250 million. This includes a previously signed deal worth $45 million in early 2023, with the potential for an additional $250 million investment by the end of the agreement period.

The partnership aims leverages Intelsat’s expertise in Geostationary Orbit (GEO) and terrestrial networks alongside OneWeb’s LEO Network, enabling the delivery of advanced, multi-orbit satellite communication solutions tailored to the needs of various sectors including networks, government, and mobility.

“Over the past couple of years, Intelsat has leveraged our partnership with the Eutelsat Group to initiate our strategy of enabling multi-orbit solutions in the commercial aviation and government markets. Today’s announcement takes that partnership to a new level, enhancing our ability to offer new, true multi-orbit services and solutions across our portfolio of customers and business segments,” said Dave Wajsgras, Chief Executive Officer of Intelsat.

Eva Berneke, Chief Executive Officer of Eutelsat Group added: “This expanded partnership with Intelsat represents a strong vote of confidence in the capabilities of the OneWeb satellite constellation, today and well into the future, as well as showcasing the necessity in today’s world for major satellite operators to be able to offer multi-orbit solutions to their customers. We are also delighted to further deepen our collaborative approach with Intelsat going forward, leveraging its unparalleled experience in the key verticals of Aviation and Government.”

Read More

Keysight aligns with NVIDIA 6G Research Cloud Platform

 Keysight Technologies is collaborating on the new NVIDIA 6G Research Cloud Platform, which includes the NVIDIA Aerial Omniverse Digital Twin, an open, flexible, and interconnected framework of network emulation resources offering researchers a comprehensive tool suite on which to develop new adaptations of artificial intelligence (AI) for radio access networks (RAN). 

Keysight has committed to add its network emulation and testing solutions covering the network from end-to-end to the NVIDIA 6G Research Cloud. Keysight is creating cloud-based versions of these solutions and making them available on the research platform in a flexible and scalable manner. This will provide 6G researchers access to Keysight solutions in network realistic conditions thanks to the NVIDIA 6G Research Cloud. As cloud-based solutions, the Keysight network emulations will now also be faster and more scalable thanks to access to NVIDIA’s seamless GPU-acceleration network.

The addition of Keysight’s portfolio of network emulation solutions will enable researchers to develop and validate new approaches to optimize wireless communications access using Keysight’s full suite of end-to-end network emulation capabilities.

https://www.keysight.com

Read More