From IoT to Cloud: A Network Function Virtualization end-to-end communication solution

点击翻译按钮获取中文摘要

During the last decades, the Internet has been a key factor in building the global information society and leading economy growth. The mix of technologies in the Internet, such as Cloud Computing, Network Function Virtualization (NFV), Data Analytics, and Internet of Things (IoT), provides the necessary tools to build a service-centric Future Internet Architecture. Future Internet is expected to rely on virtualization technologies, moving away from dedicated and expensive hardware middleboxes, while IoT will be one of the main sources of data generation in this communication paradigm.******However, the main bottleneck in such an environment is the performance degradation observed when moving from dedicated hardware middleboxes to softwarized solutions. General purpose servers must now accommodate a wide range of network services with different software and hardware requirements. At the same time, Internet's resounding success has also turned out to be the biggest obstacle leading to an inefficient utilization of the underlying infrastructure. The excessive amount of the generated traffic from the IoT layer and the need for novel and more targeted networking services has led to a depletion of the available physical resources.******Thus, the long-term goal of this proposal is to establish a scalable softwarized end-to-end solution that will gracefully combine different technologies to provide networking services with high Quality of Service (QoS) and low communication cost. This includes the creation of virtual network overlays on top of the existing physical infrastructure that will interconnect the IoT layer with the Cloud.******To achieve that, first we need to find the optimal software and hardware configuration of the servers that will host the virtualized network services to optimize the attained performance in terms of throughput and delay providing a cost efficient and attractive solution. This will help address the performance degradation noticed when using softwarized networking solutions. Secondly, taking into consideration the limited resources available, especially at the network edge, appropriate allocation algorithms will be proposed to minimize the deployment cost of the services and further improve the throughput and delay performance. Finally, data analytics will be applied to the IoT traffic to remove redundant data and unravel traffic patterns to better orchestrate the network design and resource planning of the softwarized networking solutions and improve the services delivered.******By providing a network paradigm where the main Future Internet technologies will be intermixed will pave the way for concrete solutions to a promising but still vague communication model, harmonically combining theoretical and practical studies. Furthermore, the proposed program will engage PhD and MSc students that will be trained in state-of-the art research topics and equipment, preparing high-skilled personnel for the Canadian academic and industry workplace.