Sanjay Uppal presents edge AI and the software-defined edge
Edge use cases have moved into the mainstream. Enterprises are now managing hundreds and thousands of operational technology (OT) devices at edge locations. Communications service providers (CSPs) connect enterprise edge resources, boosting their offerings with fixed wireless access and driving innovation through open APIs. AI is helping drive business outcomes across distributed environments, but solutions are siloed or inconsistent. Organizations are discovering a common thread to help them manage their edge environments with intelligence: the software-defined edge.
At VMware Explore 2024 Las Vegas, Sanjay Uppal, VP and GM of Broadcom’s Software-Defined Edge division, will lead the session When the Edge Becomes the Mainstream: What’s Next in Software-Defined Edge AI. Learn how edge AI will boost your business, and how Broadcom delivers a comprehensive software-defined edge stack that our customers and partners depend on for both current and future IT and OT success.
What is edge AI?
Everyone by now has heard of AI, and many companies are getting started with generative AI projects. But what is edge AI, and how is it relevant to your business? Edge AI helps organizations consume AI at the edge, getting real business outcomes from AI-driven apps.
Gartner estimates that more than half of all enterprise data is created and consumed at the edge. There is a tremendous need and opportunity both to support AI workloads on the infrastructure where these apps run, and to build AI into that infrastructure so that it understands and responds to the needs of the workloads.
In edge AI, data and algorithms are processed directly on devices at the edge. Edge workloads have very specific operational requirements. For example, an edge workload can’t afford the latency and processing delay of transiting through a data center, or it must run in a disconnected environment because of production, location, or sovereignty reasons.
Imagine AI-powered robots on a precisely choreographed manufacturing line, or a camera that uses computer vision with AI to recognize potential theft in a retail store. Here the workload is not a data set or a file being transmitted. The device and the workload are essentially the same thing. If the workload—the robot, the camera—can’t get the right resources at the right time, then the manufacturing line gets jammed. The store loses goods. The business suffers.
Software-defined edge is the solution
Edge AI workloads require an equally intelligent infrastructure that orchestrates the resources they need to get the job done. In the growing world of edge AI, discrete workloads administer themselves—they don’t need an “IT administrator.” In the software-defined edge model, the workload itself defines the resources it needs such as compute power, connectivity type or speed, or the maximum latency it can tolerate. The software-defined edge understands what kind of workloads are running on it, has the intelligence to communicate the needs of the workload to compute and network resources, and has the ability to orchestrate those resources for optimal results. This model can bring unparalleled efficiency and intelligence to any organization at the edge.
Sanjay Uppal will also show you the importance of fixed wireless access (FWA) to edge AI and the software-defined edge. The FWA market has made an evolutionary leap in the past year and is now more than merely a backup connectivity method for remote locations without access to broadband. With our unique approach to good/better/best enterprise connectivity at the edge, organizations can enjoy seamless broadband, cellular, and orbital availability and performance.
Featured guest speakers
We are pleased to welcome special guests who will speak with Sanjay Uppal on the impact of edge AI and the infrastructure you need to make edge AI work:
Bill Chang, CEO, Digital InfraCo, Singtel Group and CEO, Nxera
View the content catalog and secure your seats now for the sessions you want before they fill up. Make sure to click the “add to schedule” button (clicking the heart icon saves a session as a favorite but does not add it to your schedule.)
