Co-Authors: Sridhar Kayathi, Intel; Naga Rayapati, Guise AI; Lindi Sabloff, Guise AI; John Archer, Red Hat; and Darek Fanton, OnLogic
AI Is Racing to the Edge
The explosion of connected devices and the need for real-time data processing continue to drive computing from the data center to the edge at a record pace. Edge AI is now joining in, redefining how businesses operate, allowing them to make decisions efficiently and incredibly quickly.
As computational power moves closer and closer to the data source, edge AI reduces reliance on connectivity, improves latency, clears bandwidth bottlenecks, and addresses security. This opens countless possibilities, from energy forecasting to predictive maintenance in manufacturing, even AI-powered instruments in healthcare, and more.
How Does it Work?
Like everything, it starts with the need for the right tools, preferably something turnkey that can easily and seamlessly integrate into your existing operations and meet your individual needs. You won’t be surprised when I tell you we have a great solution.
Edge devices such as IoT gateways, assembly lines, points of sale/information, and vehicle systems must operate with limited computing resources, power, cooling, and connectivity. Adding to that, they are often hard to access, with little or no on-site technical expertise. You’ll need a flexible platform that can consistently support different workloads across devices and locations to meet these demands.
Core Computing + OpenVINO™
Intel Core and Intel Xeon Scalable processors with built-in AI accelerators (AMX) are the processors pushing the boundaries for Edge AI. OpenVINO™ is the free, open-source toolkit made to help you optimize a deep learning model from a framework and deploy using an inference engine. By embracing open-source software libraries and APIs powered by oneAPI, Intel empowers its partners to innovate faster while pushing the limits of edge AI applications.
Guise AI
Guise AI is a software company at the leading “edge” of AI (so to speak), solving real-world problems for some of the world’s most critical industries, including manufacturing, energy, and logistics.
For Red Hat Summit 2023, Guise AI built an “automated visual inspection” demonstration. They achieved a 62% increase in the number of frames processed per second using OpenVINO™ to meet real-time inference needs vs. TFLite, as illustrated below.
Fig 1: Intel does not control or audit third-party data. You should consult other sources to evaluate accuracy.
Guise AI solutions leverage the advantages of the device edge. Their proprietary edge AI-enabled solutions run where data is generated, offering rapid response times with low latency, high privacy and security, lower data transfer costs, and efficient use of network bandwidth.
Scaling the edge requires a comprehensive edge management platform. Guise EdgeOps is a no-code platform built to manage devices and deploy, orchestrate, and manage AI workloads at the edge in a completely secure environment. AI is also about iteration; Guise EdgeOps collects the inference and drift data securely from edge devices, building a Unified Data Store that facilitates the ability to retrain models at the edge and enables a rich hybrid cloud. The Unified Data Store allows an enterprise to choose its data gravity and enables greater “speed to ROI” with faster POCs and increasingly accurate models in production.
Red Hat
Red Hat Device Edge delivers an enterprise-ready and supported distribution of the Red Hat-led open-source community project MicroShift, a lightweight Kubernetes orchestration solution built from the edge capabilities of Red Hat OpenShift, along with an edge-optimized operating system built from Red Hat Enterprise Linux. Customers leveraging Red Hat Device Edge can achieve operational and development consistency across edge and hybrid cloud environments, regardless of where devices are deployed.
OnLogic
Edge AI solutions are frequently deployed in locations that would challenge, or even destroy, traditional computer hardware and where traditional server hardware is too big or too hot. That's where industrial computers from OnLogic come in. OnLogic is a global computer manufacturer that designs highly configurable, solution-focused systems engineered for reliability at the edge. Operating in the world’s harshest environments, they empower customers to solve complex AI implementation challenges, no matter their industry.
Bringing it All Together
Here's how enterprises can create their own optimal setup:
- Leverage Guise AI proprietary edge workloads optimized with OpenVINO™ toolkit;
- Deploy and manage edge use cases using the Guise EdgeOps platform;
- Capitalize on the power and ease of use of Red Hat Device Edge;
- Then, run it all on Intel processors on OnLogic industrial computers.
The Result
Create a trusted supply chain, reducing risk while leveraging the latest technologies to address the current realities of your business.
AI is no longer aspirational; it's accessible. Intel, Red Hat, OnLogic, and Guise AI have built a flexible Edge AI solution to solve your edge needs across any industry.
Learn More about the collaboration between Intel, Red Hat, Guise AI and OnLogic
Learn More about AI and Machine Learning
Josh Hilliker is a Senior Cloud Solutions Architect in Data Center Sales focused on advancing our Autonomous Data Center solutions, including telemetry and infrastructure modernization. Josh has held many past roles at Intel including the following: Director in SMG University, Platform Marketing in Business Client Platform Division, Intel GE-Care Innovations as the Director of Social Media, Online Presence and Google Analytics for consumer products, Product Engineer with the IT Land Capital and Acquisitions team, Operations manager for Intel e-Business Infrastructure & OEM Business Link, Architect / Community manager for Intel® vPro ™ Expert Center, Intel Manufacturing, and Director of Small Business initiatives in Intel Architecture Group.