Adlink Technology and Charles Industries Ltd. have joined together to demonstrate the industry’s first complete micro-edge, low-latency, AI/machine learning/deep learning (AI/ML/DL) solution that can be co-located on LTE small cell poles or with emerging 5G radios.

The new solution is a compact, low-profile, pole or wall-mountable unit based on an integration of Adlink’s latest AI Edge Server MECS-7210 and Charles’ SC102 Micro Edge Enclosure. The companies say this advanced solution is ideal for outdoor telecom uses and well-positioned to be deployed globally, enabling carriers and communities to provide a wide range of new and advanced services, including autonomous vehicles/pods, VR and AR applications and vision analytics with additional AI/ML/DL applications.

The solution is powered by advanced, dual-socket Intel Xeon Scalable processors and designed for all I/O front access, dust prevention, wide operation temperature and dual full-height, full-length (FHFL) PCIe expansion slots reserved for access to acceleration hardware (FGPA/GPU). Adlink’s MECS-7210 is among the first platforms to fully comply with Open Telecom IT Infrastructure (OTII) defined by Open Data Center Committee (ODCC) to meet 5G requirements of ultra-low latency, high bandwidth and real-time access to radio networks.

Charles’ SC102 Micro-Edge Enclosure has been custom-designed for the MECS-7210, using advanced thermal modeling to ensure reliable, high-end AI and computer equipment performance in harsh outdoor operating environments. The SC102 Micro-Edge Enclosures complies with the stringent Bellcore GR-487 specifications for an outdoor ambient environment of -40° C to 55° C.

The enclosed solution integrates system cooling and power by incorporating a 1,880 W heat exchanger and 250 W heater, and providing power feed options for either AC or DC. In addition, the enclosure is sealed against environmental elements such as rain, wind, ice and sun and equipped with alarms for temperature and door intrusion. The MEC-7210 server and SC102 enclosure are available as an integrated unit as well as separately.

The solution is being demonstrated at the Nvidia GPU Technology Conference (GTC), March 19 to 21, at the San Jose McEnery Convention Center in Booth 317.