Design Of A Fault Tolerance Scheme For Multi-homing In Host Identity Protocol

A host with multiple IP addresses is referred as multi-homed host. With the great development of communication technology, advanced hosts, such as mobile phones and laptops can access Internet by different technology simultaneously like WLAN and 3G. Multi-homed host has a noticeable advantage:fault tolerance. With the fault tolerance capability, a multi-homed host can switch to other link if current used link breaks down. Fault tolerance enhances the reliability of communication. But unfortunately, the primitive design principle of TCP/IP based Internet can not meet the requirements of multi-homed host. TCP/IP has manifested its primary shortcoming which is the close-coupling of transport layer and network layer. IP address is used both for identifying the identity of the host and describing the topological location. HIP (Host Identity Protocol) separates the host identity property from dual roles of IP addresses through introducing a new namespace called the Host Identity namespace. Nowadays, research work based on HIP has become the hot topics among the many Identity/Locator split schemes.In this paper, we studied the fault tolerance of multi-homed host in HIP to provide efficient and reliable support:The thesis firstly introduced the basic principle and process of Host Identity Protocol, and analyzed its mechanism for multi-homing. HIP supports multi-homing by map the multiple IP addresses of a multi-homed host to a same identity in Host Identity namespace, so that applications can use the identity for communication. Based on the implementation ideas of multi-homing, we describe a fault tolerance scheme that can provide link monitoring mechanism and failure recovery mechanism to achieve fault tolerance for multi-homed host. The thesis also presented the design and programming procedure of a fault tolerance simulation framework based on HIPSim++. To analyze the performance of designed scheme, several groups of simulation experiment is done in OMNet++ environment. The simulation results showed that the designed scheme can achieve fault tolerance capability for multi-homed host. In the end, the summary of whole thesis and expectations for future work is given.