Session
Weekday Session 8: Advanced Technologies I
Location
Utah State University, Logan, UT
Abstract
This paper explores constellation-scale architectural trades, highlights dominant factors, and presents a decision framework for migrating or sharing mission compute functionality between ground and space segments. Over recent decades, sophisticated logic has been developed for scheduling and tasking of space assets, as well as processing and exploitation of satellite data, and this software has been traditionally hosted in ground computing. Current efforts exist to migrate this software to ground cloud-based services. The option and motivation to host some of this logic “at the edge” within the space segment has arisen as space assets are proliferated, are interlinked via transport networks, and are networked with multi-domain assets. Examples include edge-based Battle Management, Command, Control, and Communications (BMC3) being developed by the Space Development Agency and future onboard computing for commercial constellations.
Edge computing pushes workload, computation, and storage closer to data sources and onto devices at the edge of the network. Potential benefits of edge computing include increased speed of response, system reliability, robustness to disrupted networks, and data security. Yet, space-based edge nodes have disadvantages including power and mass limitations, constant physical motion, difficulty of physical access, and potential vulnerability to attacks.
This paper presents a structured decision framework with justifying rationale to provide insights and begin to address a key question of what mission compute functionality should be allocated to the space-based "edge", and under what mission or architectural conditions, versus to conventional ground-based systems. The challenge is to identify the Pareto-dominant trades and impacts to mission success. This framework will not exhaustively address all missions, architectures, and CONOPs, however it is intended to provide generalized guidelines and heuristics to support architectural decision-making. Via effects-based simulation and analysis, a set of hypotheses about ground- and edge-based architectures are evaluated and summarized along with prior research. Results for a set of key metrics and decision drivers show that edge computing for specific functionality is quantitatively valuable, especially for interoperable, multi-domain, collaborative assets.
A Decision Framework for Allocation of Constellation-Scale Mission Compute Functionality to Ground and Edge Computing
Utah State University, Logan, UT
This paper explores constellation-scale architectural trades, highlights dominant factors, and presents a decision framework for migrating or sharing mission compute functionality between ground and space segments. Over recent decades, sophisticated logic has been developed for scheduling and tasking of space assets, as well as processing and exploitation of satellite data, and this software has been traditionally hosted in ground computing. Current efforts exist to migrate this software to ground cloud-based services. The option and motivation to host some of this logic “at the edge” within the space segment has arisen as space assets are proliferated, are interlinked via transport networks, and are networked with multi-domain assets. Examples include edge-based Battle Management, Command, Control, and Communications (BMC3) being developed by the Space Development Agency and future onboard computing for commercial constellations.
Edge computing pushes workload, computation, and storage closer to data sources and onto devices at the edge of the network. Potential benefits of edge computing include increased speed of response, system reliability, robustness to disrupted networks, and data security. Yet, space-based edge nodes have disadvantages including power and mass limitations, constant physical motion, difficulty of physical access, and potential vulnerability to attacks.
This paper presents a structured decision framework with justifying rationale to provide insights and begin to address a key question of what mission compute functionality should be allocated to the space-based "edge", and under what mission or architectural conditions, versus to conventional ground-based systems. The challenge is to identify the Pareto-dominant trades and impacts to mission success. This framework will not exhaustively address all missions, architectures, and CONOPs, however it is intended to provide generalized guidelines and heuristics to support architectural decision-making. Via effects-based simulation and analysis, a set of hypotheses about ground- and edge-based architectures are evaluated and summarized along with prior research. Results for a set of key metrics and decision drivers show that edge computing for specific functionality is quantitatively valuable, especially for interoperable, multi-domain, collaborative assets.
