Artificial Intelligence | News | Insights | AiThority
[bsfp-cryptocurrency style=”widget-18″ align=”marquee” columns=”6″ coins=”selected” coins-count=”6″ coins-selected=”BTC,ETH,XRP,LTC,EOS,ADA,XLM,NEO,LTC,EOS,XEM,DASH,USDT,BNB,QTUM,XVG,ONT,ZEC,STEEM” currency=”USD” title=”Cryptocurrency Widget” show_title=”0″ icon=”” scheme=”light” bs-show-desktop=”1″ bs-show-tablet=”1″ bs-show-phone=”1″ custom-css-class=”” custom-id=”” css=”.vc_custom_1523079266073{margin-bottom: 0px !important;padding-top: 0px !important;padding-bottom: 0px !important;}”]

Software Defined Infrastructure Enabled by PCIe/CXL over Photonics for AI/ML and HPC Workloads

Drut Technologies, the pioneer in PCIe/CXL over Photonics, announced that it will team with Supermicro at Super Compute 2022 in Dallas, Texas to demonstrate a low-latency photonic fabric based total solution for enabling rack scale hyper-elastic disaggregated resource pooling.

Latest Aithority Insights: Got It AI Announces AutoFlows, a Breakthrough Autonomous Conversational AI

Drut Technologies is integrating their software defined photonic interconnect resource platform for the data center with Supermicro’s Rack Scale Solutions to provide customers with:

  • The ability to decompose data center resources into resource pools: GPUs, CPUs, Memory, Accelerators, Flash, Specialized Processors, FPGAs and sNICs
  • Resource pools attached to a low-latency photonic fabric
  • Unified software manager across resource pools allowing for the fidelity of machine creation matching workloads to resources
  • Photonic scale resource pooling
  • Integration with deployed network systems with no change/impact to network silo
  • Elegant connectivity replacing bundles of cables with fiber solving the distance and latency challenge in the data center

The integration of Drut Technologies and Supermicro Rack Scale Solutions will deliver unparalleled operational and business value for users deploying GPUs for AI/ML and HPC applications. Drut’s design approach enables the mixing and matching of GPUs and other PCIe based resources. In addition, software orchestration enables creation of dynamic machines from connected resource pools. This design approach moves away from using dedicated ethernet or InfiniBand networks for resource pooling in favor of a photonic fabric. In the case of GPUs, they can be in different racks or rows or rooms or geo-locations and allocated to machines as needed via software. GPU processing power can be independently added and upgraded separate from machine dependency. Users can have racks of GPUs independent of CPU requirements.

Related Posts
1 of 41,162

AI and ML News: An Investment Into Artificial Intelligence as Daktela Buys Coworkers.ai

“We are excited to announce this collaborative approach with Supermicro,” said William Koss, CEO of Drut. “This collaboration provides customers with the ability to utilize Supermicro’s Rack Scale  Solutions and Drut’s PCIe over Photonics architecture to bring turn-key solutions to the market today.”

We will also be demonstrating our Platform Software providing a fleet of microservices that expose Redfish compliant northbound rest APIs and is deployed on a Kubernetes cluster running on Supermicro Rack Scale Solutions. This is an Open-Source based, on-prem Hypervisor providing bare metal-as-a-service (MAAS) to provide cloud-like automation for end-users utilizing OpenStack, K8s, CEPH and Lustre. Combined with Supermico’s Hyperscale Infrastructure Management Software, users are provided with a rich source of telemetry, analytics, life-cycle management, and a unified dashboard to administer software-defined data centers at hyper scale.

Red Hat and IBM Research :  Red Hat and IBM Research Advance IT Automation with AI-Powered Capabilities for Ansible

[To share your insights with us, please write to sghosh@martechseries.com] 

Comments are closed.