ABSTRACT:

Future Wireless Networks (FWNs) will be a convergence of many fixed and mobile networking technologies including cellular, wireless LANs, and traditional wired networks. This united ubiquitous network will consist of billions of network-able devices with different networking interfaces. A common networking protocol is required to communicate among these devices and interfaces; System Architecture Evolution (SAE) documents state that Internet Protocol (IP), world-widely used in the current Internet, is likely to become that common protocol. However, traditional IP architecture has faced several known challenges, such as mobility, multihoming, privacy, path preference selection, etc., which should be resolved in FWNs. One of the difficulties in the current IP architecture is the overloading of IP addresses used both as the identity and the location of IP devices. In this paper, we propose a virtualization concept for network-able components, or (virtual) objects, which generalizes all abstract components to potentially be used in FWNs. In addition, we have explicitly separated the functions of the virtual object identity from the virtual object location (using the ID/locator split concept). The end-to-end communication is a concatenation of the involved components, called a channel. To help support the ownership and policy enforcement for trusted vs. untrusted networks, a set of (virtual) network-able components with the same interest, called a realm, is formed in a multi-tier structure. The individual policy can be enforced for each individual group of (virtual) objects and/or channels. This virtualization architecture concept, characterized by the ID/locator split concept, is well-suited for FWNs and helps eliminate problems in the current Internet.

Complete paper in Adobe Acrobat format.