Abstract

In the past century, numerous iterations of automation have changed our society significantly. In that perspective, the professional and personal availability of computing devices interconnected through the Internet has changed the way we eat, live and treat each other. Today, the Internet is a service as crucial to our society as public access to electricity, gas and water supplies. Due to its successful adoption, the Internet now serves applications that were unthinkable at the time of its initial designs when social media, online global market places and video streaming were still far out of reasonable imaginary reach. Early research initiatives worked on realizing a global network of interconnected computers, an aim clearly realized by the successful implementation of the Internet and the fact that the infrastructure still suffices to provide connectivity to an unforeseen growth and change in usage. The research field of future Internet aims at long-term improvements of the Internet architecture, trying to improve the network infrastructure such that it will also facilitate future growth and applications. In this dissertation, we have contributed to the field of future Internet by proposing, implementing and evaluating infrastructure improvements. Most of our work revolves around Software-Defined Networking (SDN), a network management architecture aiming at logical centralization and softwarization of network control through the separation of data plane and control plane functionality. In particular, we have assessed the feasibility and accuracy of network monitoring through SDN (see chapter 3), as well as contributed to the robustness and recovery of such networks under topology failure by speeding up failure detection and recovery (see chapter 4) and precomputation of network-wide per-failure protection paths (see chapter 5). In addition to SDN, we have contributed to Information-Centric Networking (ICN), a network architecture optimizing content distribution by implementing network-layer forwarding techniques and cache-placement strategies based on content identifiers. We have contributed to this field by introducing a globally-accessible namespace maintaining a feasible global-routing-table size through separation and translation of context-related and location-aggregated name components (see chapter 6). Considering the same demand for centralization and softwarization of network control found in SDN applies to other network architectures, we have designed a protocol-agnostic SDN scheme enabling fine-grained control of application-specific forwarding schemes. With our prototype, we evaluate an implementation of such an SDN-controlled ICN, demonstrating correct functionality in both partial and fully upgraded networks (see chapter 7). Besides working on future Internet topics, we have also taken a step aside and looked at more recent Internet architecture improvements. Specifically, we have performed measurements on the Domain Name System’s Security Extensions (DNSSEC). From these measurements we provide insight into the level of implementation and correctness of DNSSEC configuration. Through categorization of errors we explain their main causes and find the common denominators in misconfiguration (see chapter 8).

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call