Google LLC today announced the general availability of Local Home SDK, a development toolkit that will help connected-device makers improve the reliability and responsiveness of their products.

Connected devices such as smart kitchen appliances often offload processing tasks to a cloud-based backend. That helps manufacturers keep the complexity and cost of their devices low, but it comes with certain downsides. Devices can’t access the cloud when there’s no internet connection and, when there is, sending commands to a remote data center for processing creates network latency that slows down the user experience.

Local Home SDK reduces connected devices’ reliance on the cloud. It allows hardware makers to shift processing tasks from their gadgets to the Google Assistant smart speakers and smart displays in the users’ homes. 

Local Home SDK makes it possible to run a JavaScript-based “local fulfillment app” on the target Google Assistant device that can take on some of the work usually handled by a cloud backend. Because data only has to cross the local network, there’s less latency. The toolkit also enables developers to ensure that critical device features remain available even in the absence of an internet connection. 

“As part of the Smart Home platform, local fulfillment extends your smart home Action and routes commands to devices through the local network, benefiting users with reduced latency and higher reliability,” Google developer advocate Dave Smith wrote in a blog post today. “If a local path cannot be successfully established, commands fall back to your cloud fulfillment.”

Consumers won’t need to enable the feature on their devices specifically, but hardware makers do need approval from Google before they can release local fulfillment apps. An app running on a Google Assistant device can link to a hardware maker’s gadget using the popular mDNS, UDP or UPnP protocols, then send commands in the form of standard packets.

Local Home SDK has the potential to shift yet more processing to the edge of the network. Thanks to recent advancements in chip design, a growing number of computational tasks that once had to be performed in the cloud can now be done locally on relatively simple endpoints. Startups such as newly launched Perceive Inc. are working to accelerate the trend by developing smaller and more efficient processors for connected devices. 

Photo: Google