If you have any problems related to the accessibility of any content (or if you want to request that a specific publication be accessible), please contact us at email@example.com.
Contract Routing Architecture
AuthorKaraoglu, Hasan T.
Computer Science and Engineering
AltmetricsView Usage Statistics
The current Internet architecture has been principally designed for overcoming the interoperability issues between previously disconnected network islands. Since the focus was to ensure openness and connectivity, the architecture maintained simplistic core functionalities, i.e., the set of the capabilities that have to be compatible among the Internet stakeholders. Routing, which is the negotiation of end-to-end data paths, has been successfully handled so far, since the only expectation from the Internet was to provide the basic end-to-end connectivity. Today, however, customers are not only looking for a connectivity that just works but also supports a plethora of applications that require low-latency, high-bandwidth, and secure connections to the rest of theworld. Such value-added services cannot be realized effectively mainly because the Internet core is economically rigid and only allows bilateral, single-metric negotiation of paths in a coarse-grained manner. The revenue sharing model built on this rigid economics is unhealthy and incapable of addressing the long-term sustainability and innovation at all parts (i.e., core, edge, and end) of the Internet. To overcome these limitations, we propose Contract Routing Architecture that enables users and Internet Service Providers (ISPs) i) to express their perceived economic value of offered services, and ii) to negotiate end-to-end quality of paths with other stakeholders ina multi-lateral manner considering multiple performance metrics. In this thesis, we describe how multi-metric, multi-hop path negotiations can be achieved in a scalable manner while making economics an inherent part of routing process itself so as to amend broken revenue model of the current Internet.