Zhang Si-dong

Modeling in hierarchical mobile IPv6 and intelligent mobility management scheme

Decreasing communication and management cost is a key issue of the research of Internet mobility management. Hierarchical mobile IPv6 (HMIPv6) has been proposed to reduce the number of location registration messages in backbone network by using hierarchy agent, but resources consumption inside the hierarchical domain is increased as expense. In this paper, a concept of integrated optimization is proposed, in which network cost and bandwidth consumption of delivering management messages and data payload is optimized both inside and outside hierarchical domain in a certain degree. Mathematic models of HMIPv6 are built on the viewpoint of entire network resource, and the judgment rule is obtained in result, which determines when HMIPv6 is adopted preferably. Based on the rule, intelligent mobility management scheme (IMMS) is designed, which allows a mobile node to select a suitable mobility management mechanism from mobile IPv6 and HMIPv6 according to its working parameters. Only mobile nodes protocol stack needs to be changed in IMMS and co-operation ability is assured.Decreasing communication and management cost is a key issue of the research of Internet mobility management. Hierarchical mobile IPv6 (HMIPv6) has been proposed to reduce the number of location registration messages in backbone network by using hierarchy agent, but resources consumption inside the hierarchical domain is increased as expense. In this paper, a concept of integrated optimization is proposed, in which network cost and bandwidth consumption of delivering management messages and data payload is optimized both inside and outside hierarchical domain in a certain degree. Mathematic models of HMIPv6 are built on the viewpoint of entire network resource, and the judgment rule is obtained in result, which determines when HMIPv6 is adopted preferably. Based on the rule, intelligent mobility management scheme (IMMS) is designed, which allows a mobile node to select a suitable mobility management mechanism from mobile IPv6 and HMIPv6 according to its working parameters. Only mobile nodes protocol stack needs to be changed in IMMS and co-operation ability is assured.

Throughput modeling analysis of IEEE 802.11 DCF mechanism in multi-hop non-saturated wireless ad-hoc networks

Medium access control (MAC) protocols have a large impact on the system performance for wireless ad hoc networks. Because of the limitations of existing analytical models for ad hoc networks, many researchers have to study the impact of MAC protocols via discrete simulations. In this paper, in order to improve this disadvantage, we provided an analytical model to compute of 802.11 DCF wireless multi-hop networkspsila throughput, with the assumption of finite number of terminals. This analytical model considers wireless network MAC layer in multi-hop non-saturated situation and also quantitatively computes the function of packet discard ratio factor. The simulation result shows that even in multi-hop situation, this network throughput model is still matching well with the actual wireless Ad-Hoc networks.

Fairness evaluation of the default highspeed TCP in common operating systems

Since the TCP protocol implements in the internet, the AIMD algorithm has always been the dominating congest control algorithm for about 20 years, and most of the common operating systems follow the recommendation of the IETF to adopt the AIMD algorithm in their TCP variations. But in recent years, many high-speed congest control algorithms have been proposed for fast data transfer over high speed and long distance networks, and many operating systems take the high-speed TCP variations as default: the Linux kernel uses CUBIC TCP as default since version 2.6.18, while Compound TCP becomes default TCP in Windows Server 2008 and a part of Microsoft Windows Vista, where it is easy to be enabled. As there are still not sufficient and comprehensive performance comparison of the high-speed TCP variations, the deployment of them as default TCP in the operating systems raises concerns about the future of congest control in the internet and worries about the fairness problem. This paper is an attempt to evaluate the fairness of the default TCP in the newest Linux and Microsoft Windows. The simulation results indicate that between CUBIC TCP and Compound TCP it indeed exhibits unfairness. The CUBIC TCP is too aggressive to the Compound TCP in the high bandwidth-delay product network, and the performance of Compound TCP is unacceptable comparing to CUBIC TCP.

Research on a multihomed terminal based dynamic link management algorithm in next generation internet

With the development of the network technology, the multihomed user can access the next generation internet by several access modes, such as wire and wireless technologies. According to this situation, the link management is presented as an important issue. This paper proposes a multihomed terminal based Dynamic Link Management Algorithm (DLMA), and introduces the basic principles and implement procedure of the algorithm. Moreover, the DLMA is used to compare with request-based and quantum-based methods. DLMA is a terminal based link management algorithm which can allocate and release the link resource according to the user requirement. The advantage is to reduce the overhead of link operation while fulfilling the user demand. By the experiment analysis, we evaluate the effectiveness of the DLMA.

A QoS routing algorithm based on service resource provision balance between applications and networks

To control the dynamics in QoS routing, a new concept, QoS fluctuation function (QFF) is introduced. QFF is a tool to compute the distribution state of network service resources, such as bandwidth and buffer. Also, the source QoS routing algorithm is generalized. The new QoS routing algorithm is equipped with mechanisms for QoS monitoring. Simulation experiments show that the new algorithms support stabilizing QoS and controlling the network overload, etc.

Heuristic QoS routing based on possibility distribution and soft computing

The quality of service (QoS) is often degraded by the inefficient management of networks service resources since the QoS routing decisions are often forced to base on the imprecise state information of the resources. The impreciseness is introduced by the dynamic changes of networks and the non-negligible delay in forwarding the state messages. By managing the possibility distribution of the imprecise state information with soft computing, we propose a heuristic QoS routing algorithm in this paper. The simulation results show that the algorithm can keep the QoS levels stable and routing robust. Therefore, the algorithm can be also used in the ad hoc network, wireless networks, etc. whose resource state is dominant relatively.