Return to search

New Container Architectures for Mobile, Drone, and Cloud Computing

Containers are increasingly used across many different types of computing to isolate and control apps while efficiently sharing computing resources. By using lightweight operating system virtualization, they can provide apps with a virtual computing abstraction while imposing minimal hardware requirements and a small footprint. My thesis is that new container architectures can provide additional functionality, better resource utilization, and stronger security for mobile, drone, and cloud computing. To demonstrate this, we introduce three new container architectures that enable new mobile app migration functionality, a new notion of virtual drones and efficient utilization of drone hardware, and stronger security for cloud computing by protecting containers against untrusted operating systems.

First, we introduce Flux to support multi-surface apps, apps that seamlessly run across multiple user devices, through app migration. Flux introduces two key mechanisms to overcome device heterogeneity and residual dependencies associated with app migration to enable app migration. Selective Record/Adaptive Replay to record just those device-agnostic app calls that lead to the generation of app-specific device-dependent state in services and replay them on the target. Checkpoint/Restore in Android (CRIA) to transition an app into a state in which device-specific information the app contains can be safely discarded before checkpointing and restoring the app within a containerized environment on the new device.


Second, we introduce AnDrone, a drone-as-a-service solution that makes drones accessible in the cloud. AnDrone provides a drone virtualization architecture to leverage the fact that computational costs are cheap compared to the operational and energy costs of putting a drone in the air. This enables multiple virtual drones to run simultaneously on the same physical drone at very little additional cost. To enable multiple virtual drones to run in an isolated and secure manner, each virtual drone runs its own containerized operating system instance. AnDrone introduces a new device container architecture, providing virtual drones with secure access to a full range of drone hardware devices, including sensors such as cameras and geofenced flight control.

Finally, we introduce BlackBox, a new container architecture that provides fine-grain protection of application data confidentiality and integrity without the need to trust the operating system. BlackBox introduces a container security monitor, a small trusted computing base that creates separate and independent physical address spaces for each container, such that there is no direct information flow from container to operating system or other container physical address spaces. Containerized apps do not need to be modified, can still make full use of operating system services via system calls, yet their CPU and memory state are isolated and protected from other containers and the operating system.

Identiferoai:union.ndltd.org:columbia.edu/oai:academiccommons.columbia.edu:10.7916/e5e4-r030
Date January 2023
CreatorsVan't Hof, Alexander Edward
Source SetsColumbia University
LanguageEnglish
Detected LanguageEnglish
TypeTheses

Page generated in 0.0019 seconds