Predictable Isolation and Communication Support for Embedded and Real-Time Systems

The twin goal to minimize size, weight and power (SWaP), and to provide high functionality, is particularly challenging to the traditional embedded system software. In traditional embedded systems, isolation of mission-critical computation from untrusted software, while enabling sharing of functionality and multi-core resources is challenging. For example, in autonomous vehicles where controllers, planners, sensor fusion, telemetry processing, cloud communication, and logging must all be orchestrated together. However, the isolation structures like virtualization exhibit significant scheduling and communication overheads, that are often prohibitive.Modern internet of things (IoT) systems including various low powered sensors and smart devices rely on processing capacity at the edge while having near real-time processing latency requirements. Serverless computing provides a new paradigm for event-driven, short-lived, and stateless executions, while allowing the systems to efficiently manage the performance and resources on the resource-constrained edge systems. However, the existing frameworks use VM- or container-based isolation and traditional system scheduling for serverless functions that are heavy-weight and impact the low latency requirements.In this thesis, different system mechanisms are introduced to investigate the possibilities of predictable isolation and communication in the embedded and real-time systems. This thesis enables efficient and predictable coordination for sharing of functionality and resources, while enabling strong spatial and temporal isolation of mission-critical computation from the untrusted software in multi-core, mixed-criticality systems. This thesis also enables efficient and predictable coordination of functionality on edge systems shared by various devices, while enabling strong, light-weight isolation for these embedded and IoT serverless workloads.

Author

• Gadepalli, Phani Kishore

Language

• en

Keyword

• Embedded systems
• Edge computing
• Real-Time systems
• Predictability and Isolation
• Mixed-criticality systems
• Serverless computing

Date created

• 2020

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Computer Science

Degree

• Ph.D.

Advisor

• Parmer, Gabriel A

Committee Member(s)

• Venkataramani, Guru Prasadh
• Narahari, Bhagirath
• Wood, Timothy W

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/td96k325t