Grand Challenges in Advanced Networking Research
Many current network research areas address complex "Grand Challenges." Addressing these Grand Challenges will ensure that the communications services and infrastructure required for the 21st century are developed and implemented. Most of iCAIR's research projects address Grand Challenges and core and foundation infrastructure issues.
One major challenge is scaling the Internet from services that support several billion individuals, the approximate current number, to services that can support an additional several billion and many billions of devices. Basic Internet technologies, such as its core protocols, have proven to be the most scaleable in the history of communications. However, continuing to progress forward will require much additional innovation.
Another general Grand Challenge is meeting the need to improve the current Internet -- to create a "better Internet," for example, by removing limitations, adding capabilities, and increasing its security and reliability. Another major challenge is migrating data services from primary ones supported primarily by Layer 3 infrastructure to multi-layer, hybrid services. Another goal is to create ubiquitous communication services that are easily accessible and always available anytime, anywhere, on any device.
Multiple new requirements are rapidly emerging from services and applications related to consumers, enterprises, IoT devices, AI/ML/DP, mixed reality, virtual worlds, and many other areas. Another driver for new generation communication services and networks is data-intensive, compute-intensive and bandwidth-intensive applications, especially by global scale data-intensive science. These extremely resource-intensive applications require large-scale, high-performance data management and transport WAN services for sophisticated workflows. Such large-scale transport services require fine-grained, sophisticated control processes, including those that empower edge services, applications, and processes to interact directly with core network resources, integrating capabilities for dynamic reconfiguration.
Research projects addressing these areas are creating a fundamentally new architecture that allows replacing traditional, fairly static network infrastructure with a new communication services foundation. This highly distributed, flexible facility can support multiple networks with different characteristics, each supporting many highly differentiated services. This capability is related to innovations creating new architecture and infrastructure designs for dynamically configurable networks with virtualized resources that are programmable at all levels. Early developments in programmable networking emerged from computational science based on distributed high-performance computing environments, such as HPC centers and Grids. These initiatives attempted to transition networks to "first-class resources" in distributed computational environments, making them as addressable and programmable as other resources. (Ref: Grid Networks: Enabling Grids With Advanced Communication Technology, Eds, F. Travostino, J. Mambretti, G. Karmous-Edwards, Wiley 2006). More recently, much progress has been made by developers for programmable networks used by hyperscale cloud providers. However, large-scale distributed computational science continues to drive innovation in advanced networking.