Text

theses

This thesis presents network protocol solutions to support advanced services in the future Internet. The Internet is increasingly becoming mobile with the number of mobile end-points far exceeding the fixed hosts. At the same time, new classes of services need to be supported with vastly different requirements than traditional end-hosts such as low power internet-of-things (IoTs) and highly mobile vehicular platforms. The TCP/IP network architecture developed with static end-hosts in mind fails to meet many of these requirements and there is a need for a fundamental rethinking of the network protocols in order to support these requirements in the future Internet.

The thesis starts with a comprehensive analysis of different emerging network access scenarios and identifies the set of requirements to support such use-cases, including basic host and network mobility, wireless link variation, disconnection, IoT data forwarding, and content and contextual delivery. It then introduces the concept of a named-object architecture which is designed from ground-up keeping such requirements in mind and presents an overview of MobilityFirst as a potential named-object architectural solution.

Chapter 3 presents an edge-aware inter-domain routing (EIR) protocol which provides new abstractions of aggregated-nodes (aNodes) and virtual-links (vLinks) for expressing network topologies and edge network properties necessary to address next-generation mobility related routing scenarios and link quality variations which are inadequately supported by the border gateway protocol (BGP) in use today. Specific use-cases addressed by EIR include emerging mobility service scenarios such as routing support for mobile networks in vehicular scenarios, multipath routing over several access networks, and anycast services from mobile devices to replicated cloud services. Simulation results for protocol overhead are presented and a proof-of-concept implementation on the ORBIT testbed is used to validate performance for selected mobility use-cases.

In Chapter 4, we propose a novel push-based inter-domain multicast that leverages on the concept of named-objects and a distributed name-resolution service to maintain large-scale multicast trees. The proposed named-object multicast (NOMA) protocol achieves improved scalability and performance over conventional protocols such as PIM-SM and MSDP by simplifying multicast tree generation and management. NOMA also handles mobility of end-users, thereby allowing them to move dynamically between networks, while being associated to a multicast group. Performance evaluation results, including comparisons with IP multicast, are given using a combination of analysis and NS-3 simulation. The results show good scalability properties along with low control overhead for medium to large multicast groups.

Chapter 5 presents qualitative and quantitative comparison of the proposed protocols to alternative name-based architectural solutions, such as content centric networking (CCN) as well as protocols evolved from IP, i.e. host identity protocol (HIP) and location identifier separation protocol (LISP).

Finally, in Chapter 6, we explore the 3GPP 5G core network architecture and propose named-object protocol solutions to improve control overhead for latency-sensitive applications such as IoTs and AR/VR utilizing the cellular access network and co-located mobile edge cloud. Large scale simulation using real-world datasets and proof-of-concept prototype show improved control overhead and latency for heterogeneous access scenarios.

ETD_9479

Ph.D.

Includes bibliographical references

by Shreyasee Mukherjee

doi:10.7282/t3-gd9g-ht24

ETD doctoral

The author owns the copyright to this work.