Savi has a long-standing history as the pioneer, innovator and inventor of technologies and products that leverage sensors to improve operational analysis, efficiency and insights.
The key to unlocking end-to-end supply chain visibility begins with the connectivity provided by IoT technology and its ability to transfer massive volumes of data across multiple networks.
Data can come from any source
sensors • EDI • GPS • ERP • weather • traffic • AIS • telematics • mobile • social media • other open sources
Real-time processing for vast volumes of data
When we first began building our solutions, we knew the immense amount of IoT data and variety of data sources that would need to be ingested, cleansed and processed. Subsequently, we chose an architecture that allows near real-time processing of vast volumes of data. As hundreds of thousands of shipping containers move through ports on cargo vessels, Savi uses millions of data points to accurately detect when a single container arrives in almost any port. Faster than the ship can radio the port, Savi has already sent downstream alerts.
entire data lifecycle
We store not only the original data, but also the entire lifecycle of the data as it flows through our system. Maintaining data provenance purity allows us to re-interpret the past, re-enriching our data with new understanding
Our data scientists can re-play history to understand and model the physical world. And they can use this understanding to create algorithms for real-time wisdom.
When new data flows in, the algorithms self-improve. And as new algorithms become available, we automatically select the best available for a given situation.
Time is Everything
Our complex event processing mirrors the knowledge of industry experts. A critical characteristic of lambda hybrid architecture is how it understands time. Time comes from GPS, data travels through cellular data networks, then edge data centers around the world, before navigating to our servers. Savi’s processing systems understand that order of receipt is not the same as order of occurrence. Thus, we have built in logic to mitigate order-dependent issues. Any step along the line could re-order data. Our event processing handles this with ease.