Planetary Scale Digital Twin
Built on a digital twin of wireless transceivers and signal propagation at planet-scale, Spacetime has a full celestial frame of reference. It can understand the physics of the Earth and space, as well as motion propagation, weather phenomena and atmospheric characteristics: everything that affects a wireless signal. It also uses the position, orientation and projected motion of the physical platforms to forecast the quality and stability of upcoming connection opportunities, enabling it to proactively evolve the network topology and increase network resilience.
Google initiates several projects to connect the unconnected using LEO satellite constellations, HAPS, and UAVs.
Google begins work on “Minkowski”, a software platform for Temporospatial Software Defined Networking (TS-SDN), in order to support the network control and orchestration of LEO satellite constellations, HAPS, and other dynamic mesh networks.
First successful test of TS-SDN with an airborne communications payload
TS-SDN deployed at scale; performs lights-out orchestration of the ground stations and communications payloads in Loon’s stratospheric mesh network
TS-SDN achieves 1 million flight hours
TS-SDN achieves 2 million flight hours
Aalyria is formed; acquires TS-SDN technology (Spacetime) from Google
Spacetime achieves General Availability on Google Cloud for SaaS or self-hosting
Spacetime is selected to orchestrate Rivada Space Networks LEO constellation
The Network Fabric of Spacetime
Spacetime blends traditional SDN / SD-WAN features for the enterprise network and ground segment with control plane orchestration of wireless networks and directional or steerable beams. This novel approach, called Temporospatial Software-Defined Networking (TS-SDN), augments the network information base with a digital twin capable of modeling the position, orientation, and motion of any physical platforms and forecasting the opportunities for wireless links and coverage opportunities over time. Leveraging predictability in motion, weather, and faster-than-realtime modeling allows Spacetime to jointly optimize and solve the steerable beam tasking and scheduling, radio and optical transceiver resource management, and path-agnostic route orchestration across time and space, in all domains.
Spacetime’s open Northbound APIs can be used to dynamically express network requirements, such as a new request to provision resilient, end-to-end network transit at a specified priority and data rate. The endpoints for such requests can be expressed referencing the true source and destination of those flows, including any combination of enterprise services, cloud resources, coverage regions, or network interfaces across land, sea, air, or space. Spacetime optimizes and orchestrates the network control plane accordingly by continuously installing, revising, and monitoring the scheduled enactment of network state transitions and telemetry across the network through its open Controller-to-Dataplane Interface APIs. All of Spacetime’s APIs are based on gRPC, are available on Github as open source, and are provided to government customers with unlimited rights.
Spacetime’s digital twin predicts wireless access, backhaul, and ISL opportunities with full consideration of near-term motion, including but not limited to aircraft bank angles, field-of-regard obstructions, ground station horizon masks, terrain, and sun outages. This enables a proactive make-before-break network flow on an end-to-end network path basis, while proactively considering wireless link capacities, interference, and detectability. NOAA weather nowcasts/forecasts are integrated into its wireless signal propagation modeling supporting proactive orchestration around atmospheric, rain, clouds, fog, and scintillation for any RF frequency up to 100 GHz as well as atmospheric free space optics.
Spacetime’s multi-domain, temporospatial digital twin supports land, sea, air, and space – including terrestrial terminals and handsets, ground stations, ships, aircraft, HAPS, LEO, MEO, and GEO satellite constellations, cislunar, and deep space network nodes. New physical topologies are proactively created, and routes are then migrated across all domains to ensure that existing network flows experience zero packet loss. This enables it to reoptimize the network, across all domains, within seconds in response to unpredictable events, such as the loss of a satellite, a cut to a submarine fiber, or a new request for a high priority network transit.
Spacetime’s interface provides situational awareness across the geospatial and logical representations of an all-domain network. Users can:
- Take actions such as requesting services, draining traffic from the network, and other operational scenarios
- View and replay historical context and state transitions to diagnose outages
- View and replay future context for analysis