King-Tim Ko

PD-RED: to improve the performance of RED

We propose a new active queue management (AQM) scheme to improve the performance of the well-known random early detection (RED) AQM. The new AQM is based on the proportional derivative (PD) control principle, and we call it PD-RED. In PD-RED we introduce minimal changes to RED. We demonstrate the improvement in performance of PD-RED over adaptive RED AQM by simulations.

Using genetic algorithms to design mesh networks

Designs for mesh communication networks must meet conflicting, interdependent requirements. This sets the stage for a complex problem with a solution that targets optimal topological connections, routing, and link capacity assignments. These assignments must minimize cost while satisfying traffic requirements and keeping network delays within permissible values. Since such a problem is NP-complete, developers must use heuristic techniques to handle the complexity and solve practical problems with a modest number of nodes. One heuristic technique, genetic algorithms, appears to be ideal to handle the design of mesh networks with capability of handling discrete values, multiobjective functions, and multiconstraint problems. Existing applications of genetic algorithms to this problem, however, have only optimized the network topology. They ignore the difficult subproblems of routing and capacity assignment, a crucial determiner of network quality and cost. This article presents a total solution to mesh network design using a genetic algorithm approach. The application is a 10-city network that links Hong Kong and nine other cities in China. The development demonstrates that this method can be used for networks of reasonable size with realistic topology and traffic requirements.

PD-controller: a new active queue management scheme

This paper describes a proportional-differential (PD) control algorithm as a new active queue management (AQM) scheme for TCP/IP congestion control. From the viewpoint of the control theory, TCP congestion control system can be regarded as a feedback regulating system. In this paper, a robust AQM called PD-controller is proposed. The design principles of PD-controller are presented in details. Its performance is extensively evaluated by simulations. The results demonstrate that the PD-controller AQM is stable and robust against traffic load fluctuations, UDP and HTTP disturbances. Its superiority over other AQMs is also demonstrated.

Approximation of loss calculation for hierarchical networks with multiservice overflows

This paper studies loss calculation in hierarchical networks with multiservice overflows which have different call arrival rates, mean holding times, bandwidth requirements and share a common link. The loss calculation involves two challenging problems: 1) the computation of the two moment characterizations of multiservice overflow traffic over the shared link, 2) the calculation of the loss probabilities for multiservice non-Poisson overflow traffic in hierarchical systems. An efficient approximation method, known as multiservice overflow approximation (MOA), is proposed to enable multiservice networks designs with hierarchical architecture. Two contributions are included in the MOA method. First, an approximation based on blocking probabilities matching is proposed to compute the variances of multiservice overflows over the shared link. Second, a modified Fredericks & Haywards approximation is used to calculate the loss probabilities of multiservice non-Poisson over flow traffic. The performance of the MOA method is evaluated in a two-tier hierarchical cellular network and compared with an existing approximation method based on multi-dimensional Markov-modulated Poisson process (MMPP). Verified by simulations, the MOA method achieves better accuracy in the general heterogeneous cases at lower computational cost than the MMPP method.

Optimal video placement scheme for batching VOD services

Advances in broadband technology are generating an increasing demand for video-on-demand (VOD) applications. In this paper, an optimal video placement scheme is proposed for a batching VOD system with multiple servers. Given a specified requirement of the blocking probability, an optimal batching interval is derived and the corresponding file placement is obtained by hybrid genetic algorithm. It is demonstrated that the specified requirement on blocking probability is satisfied, while both batching interval and server capacity usage are minimized simultaneously.

Optimal file placement in VOD system using genetic algorithm

Advances in computing and networking are generating a significant demand for video-on-demand (VOD) applications. In this paper, the condition for minimum blocking probability of a VOD system is derived. The optimal load-sharing requirement in such a system is found so that the minimum blocking probability is achieved. A combination of genetic algorithm and modified bin-packing algorithm is then developed for the multimedia file placement exercise. It is demonstrated that a specified blocking probability can be achieved while the capacity usage is minimized simultaneously.

Bandwidth allocation and routing in virtual path based ATM networks

The virtual path (VP) concept in ATM networks simplifies traffic control and resource management. For bandwidth allocation, a VP can carry traffic of the same type (the separated scheme) or of different types (the unified scheme) if the most stringent quality of service (QoS) amongst traffic of different types is used on the VP. In addition, VP capacity is dynamically adjusted in a predetermined incremental size. In this paper, we study a least loaded path-based adaptive routing algorithm called maximum free circuit routing under two bandwidth allocation schemes, in an ATM network employing the VP concept. In particular, we evaluate the call blocking probability and the call set-up processing load with varying bandwidth incremental sizes. It is found numerically how the use of VP trades the blocking probability for call setup processing load. It is also found that the unified scheme could outperform the separated scheme with the use of VP.

LAMOR: Lifetime-Aware Multipath Optimized Routing Algorithm for Video Transmission over Ad Hoc Networks

Multipath routing is a key technique to support video transmission over wireless ad hoc networks (WANETs). In WANETs, the lifetime of a node is related to its residual energy, current traffic conditions and the required energy consumed for sending a packet to its next hop in the path. In this paper, we propose a new adaptive routing scheme termed lifetime-aware multipath optimal routing (LAMOR) for supporting high-speed real time video transmission in WANETs, which is optimized in terms of lifetime and analyze its characteristics. Both theoretical analysis and simulation results demonstrate that LAMOR indeed extends network lifetime and improves the transmission quality of video streams

Buffer sizing for nonhomogeneous TCP sources

Considering a single bottleneck model of TCP with a constant number of greedy and nonhomogeneous sources and assuming that TCP timeouts do not occur, we establish necessary and sufficient conditions, related to the bandwidth delay product, to guarantee that the buffer will never empty. We also demonstrate by simulation that weaker conditions could be adequate to maintain high utilization in practice and discuss delay and packet drop rate implications.

A stability theorem for Internet congestion control

This paper proves a stability theorem for an Internet congestion control algorithm with diverse propagation delays.