The Dragonfruit Base Station is an on-premise hardware-as-a-service device running on the Apple M1 chipset, which is optimized for video and ML processing. Its core functions are bandwidth management, video ingestion into the Dragonfruit cloud, and hybrid AI processing for video intelligence.
Installing the Base Station is as easy as plugging in power and Ethernet – the configuration is then remotely finalized in the browser. Hardware cost is included in the price of the platform and it automatically connects to your existing cameras via RTSP or ONVIF.
The Base Station accommodates three main upload policies for optimizing bandwidth: Time-shifting, Down sampling, and Split Inference. Dependable cloud transport optimizes available bandwidth and TLS 1.2 encryption and access controls protect data in transit.
Upload videos based on bandwidth allocation. For example, if more bandwidth is available after-hours, video uploads will automatically shift to use more bandwidth then.
Upload videos based on quality constraints. For example, if the customer requirements call for a certain minimum resolution, and there isn’t enough bandwidth available, videos are automatically downsampled to fit the constraints.
Upload videos using ML inference. For example, if the customer is bandwidth constrained but needs deep video search, Dragonfruit will automatically deploy ML inference to balance analytics needs and infrastructure constraints.
Dragonfruit's Base Station auto-detects most camera types and automatically connects to them via RTPS/ONVIF. Low-latency video streaming allows users to see events as they're happening. A customizable video walls provides central visibility across all locations. The Base Station provides 20 camera-days of storage on the device and is easily expandable. Cost-effective cloud-storage options are also available for unlimited video retention.
Dragonfruit's Base Station is optimized for video and ML processing so much of the AI heavy-lifting can be done on the system itself. Operations such as video intelligence – people counting, dwell times, occupancy metrics, etc. – are all computed on the Base Station itself and video metadata is sent to cloud so the UI can tabulate results and users can view longitudinal data to understand what's happening in their space. For custom AI models requiring heavier processing power, Dragonfruit's Split AI is used to optimize bandwidth for video uploads.