A Consolidated Namespace for Network Applications, Developers, Administrators and Users

Miika Komu

    Research output: ThesisDoctoral ThesisCollection of Articles

    Abstract

    The current Internet is founded on the TCP/IP architecture that was originally designed around machines rather than humans. In the original architecture, computers were always named using numerical IP addresses until symbolic names were added to the architecture. Today, most end-users access Internet services directly using symbolic host names with DNS extensions or indirectly by utilizing key words with search engines, but the pervasiveness of IP addresses remains a source of inflexibility as the TCP/IP architecture. The wild success of the Internet, moreover, has attracted significant financial investments, and has resulted in a heavy financial stake in the existing infrastructure and architecture. Consequently, even conservative improvements to its ossified design can face a difficult deployment path. In this dissertation, we examine a number of legacy-compatible and evolutionary solutions to three of the challenges in the TCP/IP architecture. The first challenge of non-persistent addressing stems from the reuse of IP addresses at network, transport and application layers. While this simplified the naming model of the original TCP/IP architecture, it disrupts TCP streams when the topological location of a mobile device changes. As with other causes of non-persistent addressing, Internet transparency is lost as NAT devices are based on private address realms, and site renumbering is difficult as addresses are hard coded into various configurations. The second challenge is that heterogeneous addressing, as introduced by IPv6, complicates the addressing of hosts and the networking logic of applications. The third challenge is that the addressing model offers little support for security, which needs to be reinforced at the various layers of the networking stack. A consolidated namespace meets the requirements of these three high-level challenges. From the surveyed solutions, we have narrowed down the number of alternatives to seven solutions that fulfill the requirements of a consolidated namespace and used the Host Identity Protocol (HIP) for empirical evaluation. As other work exists in this area, our work focuses on application layer aspects because it has remained relatively unexplored, especially in the context of HIP. The concrete research problems are threefold. First, we revisit some aspects of the challenges for consolidated naming at the application layer to understand the impact of the problems. Then, we implement improvements on HIP to better meet the goals for consolidated naming for end-users, network application developers and network administrators. Thirdly, we design, develop and analyze technical improvements to HIP in order to facilitate its adoption and deployment.
    Translated title of the contributionLujitettu nimiavaruus verkkosovelluksille, kehittäjille, ylläpitäjille ja käyttäjille
    Original languageEnglish
    QualificationDoctor's degree
    Awarding Institution
    • Aalto University
    Supervisors/Advisors
    • Ylä-Jääski, Antti, Supervising Professor
    • Tarkoma, Sasu, Thesis Advisor
    • Gurtov, Andrei, Thesis Advisor
    Publisher
    Print ISBNs978-952-60-4904-5
    Electronic ISBNs978-952-60-4905-2
    Publication statusPublished - 2012
    MoE publication typeG5 Doctoral dissertation (article)

    Keywords

    • mobility
    • multihoming
    • site renumbering
    • IPv6
    • security
    • HIP

    Fingerprint

    Dive into the research topics of 'A Consolidated Namespace for Network Applications, Developers, Administrators and Users'. Together they form a unique fingerprint.

    Cite this