Recently, NXP Semiconductors and SoftAtHome announced a joint effort to develop a solution to achieve edge computing for home gateways. The solution is based on SoftAtHome's Connect'ON modular software platform and NXP's LayerscapeTM multi-core processor, uses a Linux container to host applications that can be deployed to the home gateway in real time through orchestration in the cloud.
Edge Computing technology has the power of cloud computing and artificial intelligence while meeting low latency, privacy and reliability requirements. For example, an application such as facial recognition for "video doorbell" may run machine learning in the cloud and download the inference model to the home gateway at set intervals. This means that camera data does not need to flow continuously to the cloud, and the application can continue to work even if the Internet connection is lost. NXP's EdgeScale technology for entry and management of IoT device security covers the entire data flow from the edge to the cloud, providing end-to-end security for edge computing solutions.
The edge solution approach has several advantages over traditional application architectures, including the ability to dynamically deploy without rebooting. It also securely separates applications from critical services and leverages cloud computing, edge computing capabilities and artificial intelligence capabilities. The most important innovation is that the same runtime environment can be used on cloud and edge devices such as Home Gateway, so developers can write new applications without knowing the embedded system.
"Combining SoftAtHome's modular gateway platform with NXP's EdgeScale technology and high-performance multi-core processors will accelerate the availability of new home computing models," commented Noy Kucuk, vice president of digital networks at NXP. “We are pleased to work with SoftAtHome to provide new services based on artificial intelligence in the hybrid cloud.”