Taking a hybrid approach to distributed computing with AMD Kria SOM

Latest update time：2024-08-21

Reads：

The number of sensors and connected devices at the edge continues to grow at an exponential rate every day. Analog electronic sensors connected to digital computing devices enable systems to gain situational awareness and optimize performance for high productivity and efficiency. There are multiple ways to address the processing challenges of the surge in sensor data generated at the edge:

Send all collected data to the cloud for processing

This approach increases latency and high data transmission costs.

Process all collected data at the edge (close to the analog-digital boundary)

This approach requires higher processing locally

Distribute data across the cloud and the edge

This hybrid approach achieves a balance between latency, computing power, data transmission costs, and power consumption.

A hybrid approach to distributed computing can be achieved by using a scalable, efficient, and low-power adaptable computing platform at the edge that can seamlessly connect to the cloud to transmit bidirectional data. To get users started quickly, the entire AMD Kria™ Starter Kit portfolio, including its Kria SOM, is certified to run the most popular IoT management systems today - AWS IoT Greengrass and Azure IoT Edge.

The AMD Kria Starter Kit portfolio includes out-of-the-box developer platforms for designing vision AI, robotics and industrial, motor control, and DSP applications. The AMD Kria Starter Kit makes it easy for embedded software developers without FPGA expertise to start building unique edge applications and solutions using the Kria K26 SOM for mass production deployment. Heterogeneous computing with the Kria SOM can meet edge computing requirements with low latency, low power, and determinism while assuming distributed sensor fusion and AI between the cloud and the edge. Certified for AWS IoT Greengrass and Azure IoT Edge, it has never been easier to start a design with AMD.

What are IoT management systems and why are they important?

In a typical deployment, there will be multiple edge devices connected with sensors that communicate with the cloud. IoT management systems allow users to register these edge devices to establish communication with the cloud, create groups, collect data, push updates, communicate locally with other edge devices, and keep security top of mind!

AWS IoT Greengrass is an open source IoT edge runtime and cloud service that helps you build, deploy, and manage IoT applications on devices. AWS IoT Greengrass enables your devices to collect and analyze data closer to where the data is generated, respond autonomously to local events, and communicate securely with other devices on the local network. Greengrass devices can also communicate privately with AWS IoT Core and export IoT data to the AWS Cloud.

Azure provides IoT Edge, a device-centric runtime that enables you to deploy, run, and monitor containerized Linux workloads. Azure IoT Edge is a feature of Azure IoT Hub that enables you to scale and manage IoT solutions from the cloud. Azure IoT Edge helps you bring cloud analytics closer to your devices, enabling better business insights and offline decision making.

Getting Started with the AMD Kria Starter Kit

Kria SOMs are designed with software engineers in mind, providing a familiar design environment with no FPGA programming experience required. They are supported by Kria Starter Kits, which are low-cost, out-of-the-box development platforms certified to work with AWS IoT Greengrass and Azure IoT Edge. It has never been easier to start development with AMD products and realize the benefits of distributed computing across the cloud and edge.

Learn more

For more information about AMD Kria product portfolio , please visit the official website Product section .