The “cloud paradigm” can provide a wealth of sophisticated emergency communication services that are game changers in emergency response, but the current implementation of the cloud paradigm is not suitable to the challenging environments in which these responses often take place. The networking infrastructure may be all but unavailable, and access to centralized data centers may be impossible.
What is needed is an entirely new approach to implementing cloud infrastructure. This extends not only to how communication is done, but also to so-called edge cloud infrastructure. Emergency workers may bring small and resource-constrained clusters and networking infrastructure to an emergency location. Such “edge clouds” may - in theory - provide basic cloud services with improved availability and response time to handheld devices. However, current cloud software is not optimized for this and will immediately break down. Communication services in such an environment are not likely to work well until new software for such environments is developed.
Principle Investigator: Hakim Weatherspoon
This proposal proposes three disruptive software techniques to significantly improve the state-of-the-art in edge clouds for emergency response. The first is X-Containers, a new way of running Docker containers that combines the security and extensibility of virtual machines with the convenience and efficiency of containers. X-Containers are both more secure and more efficient than containers and as flexible as virtual machines. The insight is that when running a single Linux container inside a virtual machine, the isolation layer that Linux implements becomes obsolete. With X-Containers, we can provide improved security and efficiency while also be able to leverage special purpose communication devices.
The second technique is vertical scaling and placement optimized for emergency services. The cloud only provides horizontal scaling (“elasticity”) and placement within centralized clusters. Vertical scaling allows transparent control over the amount of physical memory, the number of physical cores, or the amount of network bandwidth that a container has access to. In addition, containers and storage can be migrated to various locations in order to optimize availability and performance.
The third technique is what we call opportunistic routing. In addition to the usual name- or address-based messaging techniques which require a control plane, we want to add a complementary communication technique based on content-based routing of shared documents. Such documents can include annotated maps, whiteboards, slack communication channels, requests for information, and also monitoring and control documents used to manage the system itself. Routing of these documents will be based on content as well as geographic need, time, dependencies, and other priorities.
If successful, the potential impact of the proposed research to public safety is significant and transformative. In particular, first responders will be able to dynamically and locally provision an edge cloud. A newly provisioned edge cloud will be able to set up computing, storage, networking, and management on available resources. Importantly, X-Containers and vertical scaling approaches will enable secure and efficient use of resources. Finally, intelligent placement of resources and opportunistic routing will enable the system to tolerate intermittent networking conditions.