Graph Architecture Meets Streaming Technology
The world of signals is characterized by massive scale, complex inter-relationships, and rapid change. SignalFrame has developed a sophisticated platform to create and update graphs from streaming data. The platform delivers powerful analytics over community clusters — and does so within temporal constraints.
Our engineering team uses many open-source projects such as Kubernetes, Cassandra, Kafka, and ZooKeeper. Inspired by other open-source projects, we have built our clustering, streaming, and indexing services in Go.
The SignalGraph grows out of crowdsourced observations of the ambient signal environment that reveal the presence and proximity of wifi, bluetooth and BLE devices in the environment. Our network of more than 5mm global devices observes 500mm distinct signals per month.
SignalGraph is self-organizing, constructing relationships from the bottom up, and adapting to changes. Signal observations are combined with our semantic resolution engine to bring real world meaning to signals and clusters of signals. Streaming integration delivers low-latency access to the full graph, powering real-time proximity services and always-on push analytics.