Altran and Intel open up edge AI telecoms platform
French engineering group Altran has made the first key steps with Intel on an open edge AI platform for telecoms operators that will launch across Europe later this year..
Altran, now part of Capgemini, is working with a consortium of network operators, technology vendors and the GSM Association to create the GSMA Operator Platform. This will allow operators to cooperatively provide edge services, regardless of location and the platform will be shown later this year as a proof of concept.
This multi-access edge computing (MEC) system will be based on Altran’s ENSCONCE software combined with Intel’s Open Network Edge Services Software (OpenNESS), along with other Intel edge AI hardware and software technologies.
ENSCONCE sits on micro data centres close to the access network, aggregation points, regional data centres and central offices to host edge applications. The platform supports low-latency edge application development through software development kits (SDKs). The SDK provides edge applications on demand, discovers edge deployments, orchestrates applications across operator networks, and monitors and manages applications throughout the lifecycle.
OpenNESS enables ENSCONCE to provide uniform interfaces for 4G and 5G mobile networks to reduce network latency. It also supports a library of optimized inferencing models for computer vision called OpenVINO. This executes low-latency inferencing on a variety of edge AI accelerators, including Intel’s Xeon processors, Movidius Vision Processing Units (VPUs), Intel FPGAs and neural accelerators.
“Our MEC platform is focused on developer experience and will incorporate relevant application programming interfaces for application developers in both their SDKs and for developer-facing services,” said Pascal Brier, Executive Vice-President of Strategy, Technology and Innovation at Altran. “By integrating Intel technology and toolkits including OpenNESS and OpenVINO into our platform, we will significantly amplify our edge computing solution and strengthen the offering for application developers and operators.”
Next: Edge AI collaboration
“There is a growing need for edge platforms that provide a seamless developer experience and can be implemented across a variety of use cases,” said Renu Navale, Vice-President of the Data Platforms Group and General Manager of Edge Computing & Ecosystem Enabling at Intel. “The collaboration with Altran to integrate various modules from the OpenNESS toolkit along with other Intel technology components such as the Intel Distribution of OpenVINO toolkit will result in a compelling edge platform that is developer friendly and accelerates ecosystem innovations.”
Altran was acquired by French technology consultancy Capgemini back in April and is working with China Unicom, Deutsche Telekom, EE, KDDI, Orange, Singtel, SK Telecom, Telefonica and TIM on the edge AI compute architectural framework and reference platform.
This Operator Platform Project will be deployed across multiple markets in Europe first and then extended to other operators and geographies.
“Edge computing technology is getting mature and will soon become ready to support edge applications with the performance 5G is promising. Collaborations like this are a great step forward to enable solid commercial solutions to support our future edge services,” said Juan Carlos García, Senior Vice-President of Technology and Ecosystems at Telefonica. “The joint work of Intel and Altran, both highly committed with the activity in standardization, open-source and industry bodies like ETSI, GSMA and LFN, will contribute to the availability of open, global and interoperable solutions for MEC.”
“Operators are very well placed to provide capabilities such as low latency through their network assets,” said Alex Sinclair, CTO at GSMA. “It is essential for enterprises to be able to reach all of their customers from the edge of any network. Based on the GSMA Operator Platform Specification, Telco Edge Cloud will provide enterprise developers and aggregators with a consistent way to reach connected customers.”
“Edge cloud will build a unified network edge ecosystem, providing diversified and customised products and services, and multiple platform capabilities. It will also realise more extensive boundary-crossing cooperation to meet the requirements of digital transformation of various vertical industries,” said Xiongyan Tang, the Chief Scientist of China Unicom Network Technology Research Institute and the Chief Architect of China Unicom Intelligent Network Centre.
Related edge AI articles
- LOW POWER BREAKTHROUGH FOR EDGE CHIPS
- BLAIZE DETAILS ARCHITECTURE OF GSP CHIP
- TRAINING ON MICROCONTROLLERS BOOSTS IoT PREDICTIVE MAINTENANCE
- STARTUP LAUNCHES EDGE BOARDS AND TOOLS
Other articles on eeNews Europe