Search

This work was partially funded by the Government of Thailand through a PhD scholarship for Dr Phoomikiattisak. As the use of mobile devices and methods of wireless connectivity continue to increase, seamless mobility becomes more desirable and important. The current IETF Mobile IP standard relies on additional network entities for mobility management, can have poor performance, and has seen little deployment in real networks. We present a host-based mobility solution with a true end-to-end architecture using the Identifier-Locator Network Protocol (ILNP). We show how the TCP code in the Linux kernel can be extended allowing legacy TCP applications that use the standard C sockets API to operate over ILNP without requiring changes or recompilation. Our direct testbed performance comparison shows that ILNP provides better host mobility support than Mobile IPv6 in terms of session continuity, packet loss, and handoff delay for TCP. Publisher PDF

D. Phoomikiattisak was funded by the Thai Government. B. Simpson was funded by Cisco Systems under a University Research Programme (URP) grant award. We discuss a key engineering challenge in implementing the Identifier- Locator Network Protocol (ILNP), as described in IRTF Experimental RFCs 6740-6748: enabling legacy applications that use the C sockets API. We have built the first two OS kernel implementations of ILNPv6 (ILNP as a superset of IPv6), in both the Linux OS kernel and the FreeBSD OS kernel. Our evaluation is in comparison with IPv6, in the context of a topical and challenging scenario: host mobility implemented as a purely end-to-end function. Our experiments show that ILNPv6 has excellent potential for deployment using existing IPv6 infrastructure, whilst offering the new properties and functionality of ILNP. Postprint

Seamless host mobility is vital to future network mobility, and has been an active research area for a long time. Much research focuses on the performance of the data plane. In this paper, we present comprehensive analyses on the control (signalling) plane in the IETF Mobile IPv6, and compare it with the IRTF Identifier-Locator Network Protocol (ILNP). The control plane behaviour is important in order to assess the robustness and scalability of the mobility protocol. ILNP has a different mobility model from Mobile IPv6: it isa host-based, end-to-end architecture and does not require additional network-layer entities. Hence, the control signals are exchanged only between the end systems. We provide model-based analyses for handoff signalling, and show that ILNP is more efficient than MIPv6 in terms of robustness and scalability. The analytical models we present could also be adapted for other mobility solutions, for comparative assessment. Postprint

Seamless host mobility has been a desirable feature for a long time, but was not part of the original design of the Internet architecture or protocols. Current approaches to network-layer mobility typically require additional network-layer entities for mobility management, which add complexity to the current engineering landscape of the Internet. We present a host-based, end-to-end architecture for host mobility using the Identifier-Locator Network Protocol (ILNP). ILNP provides mobility support as a first class function, since mobility management is controlled and managed by the end-systems, and does not require additional network-layer entities. We demonstrate an instance of ILNP that is a superset of IPv6 – called ILNPv6 – that is implemented by extending the current IPv6 code in the Linux kernel. We make a direct comparison of performance of ILNPv6 and Mobile IPv6, showing the improved performance of ILNPv6. Postprint

Date of Acceptance: 14/07/2014 We present the first results of an implementation of IP-layer soft handoff, based on the Identifier Locator Network Protocol (ILNP). In our testbed experiments, we show minimal gratuitous packet loss in vertical handoff scenarios (WiFi-3G). Unlike the IETF Mobile IP proposals, the ILNP uses a purely end-to-end architecture, and does not require proxies, middleboxes or tunnelling to support mobility. Our testbed is based on an in-kernel implementation using a modified Linux IP stack. Postprint