Software Defined Flexible and Dynamic Network Resource Allocation and Resilience

Lead PI: Nael Abu-Ghazaleh 

The network layer is a critical part of the system bridging the application layer and the physical layer.  It is responsible for efficient, resilient and secure allocation of the network and edge computing resources in a way that satisfies application requirements of the applications.  Our solutions will also leverage machine learning to create novel data driven resource allocation strategies and will develop new transport protocols that cater for emerging learning based applications within a system that is agile, resilient and secure.  

The research in this space is in two directions: (1) Secure and resilient data driven resource allocation; and (2) Efficient and agile network system support.  In the first direction, we start with application level resource slice demands and then reason about the available resource to efficiently support these application demands.  We also seek to integrate data driven approaches to create proactive, resilient and efficient resource allocation strategies.  We also consider the development of new network services and protocols that can support machine learning and other data driven applications where the consumer of data is a computational, rather than a human, end point.  In the second direction, we consider support of processing and networking services in a dynamic edge computing environment requires systems capable of meeting application as well as the system needs, such as executing algorithms for allocation of wireless resources.  Facets of this problem include the support of resilient software defined networking (SDN); agile Network Function Virtualization (NFV) capabilities; software virtualization for rapidly instantiating applications and migrating them on demand.