Shoji Kasahara

Practical Time-Scale Fitting of Self-Similar Traffic with Markov-Modulated Poisson Process

Recent measurements of packet/cell streams in multimedia communication networks have revealed that they have the self-similar property and are of different characteristics from traditional traffic streams. In this paper, we first give some definitions of self-similarity. Then, we propose a fitting method for the self-similar traffic in terms of Markov-modulated Poisson process (MMPP). We construct an MMPP as the superposition of two-state MMPPs and fit it so as to match the variance function over several time-scales. Numerical examples show that the variance function of the self-similar process can be well represented by that of resulting MMPPs. We also examine the queueing behavior of the resulting MMPP/D/1 queueing systems. We compare the analytical results of MMPP/D/1 with the simulation ones of the queueing system with self-similar input.

A simple algorithm for the rate matrices of level-dependent QBD processes

We develop a simple algorithm to compute the rate matrices of ergodic level-dependent quasi-birth-and-death processes, based on a matrix continued fraction representation and a probabilistic interpretation of those matrices. The algorithm is easier to implement and less memory-consuming than that developed by Bright and Taylor. We apply the algorithm to an infinite-server queueing system with a Markovian arrival process and exponentially distributed service times. We present some numerical results in order to demonstrate the efficiency of our algorithm.

Burst-cluster transmission: service differentiation mechanism for immediate reservation in optical burst switching networks

In this article we propose a service differentiated burst transmission called burst-cluster transmission to provide service differentiation in terms of burst loss for optical burst switching networks. The proposed method consists of a burst assembly algorithm and burst transmission scheduling, and it works only at the ingress edge node. A mixed timer/burst-length-based assembly algorithm is considered for burst assembly, and bursts with different service classes are assembled simultaneously according to the algorithm. A burst-cluster is generated so that the bursts in the cluster are arranged in order from lowest priority to highest. Then the burst-cluster is transmitted according to the burst transmission scheduling. We also consider the application of burst-cluster transmission to delay-sensitive traffic. We evaluate by simulation the performance of the burst-cluster transmission for NSFNET. Numerical examples show that burst-cluster transmission succeeds in providing different service grades even though the burst loss probabilities of high-priority service classes are affected by the traffic load of low-priority ones

Performance analysis of timer-based burst assembly with slotted scheduling for optical burst switching networks

In this paper, we analyze the performance of a timer-based burst assembly for optical burst switching (OBS) networks. In our analytical model, an ingress edge node has multiple buffers where IP packets are stored depending on their egress edge nodes, and bursts are assembled at the buffers in round-robin manner. Moreover, bursts are transmitted in accordance with slotted scheduling where each burst transmission starts at the slot boundary. We construct a loss model with two independent arrival streams, and explicitly derive the burst loss probability, burst throughput, and data throughput. In numerical examples, we show the effectiveness of our analysis in comparison with the Erlang loss system. It is shown that our model is quite useful for an OBS network with a large number of input and output links.

Round-robin burst assembly and constant transmission scheduling for optical burst switching networks

In this paper, we propose a round-robin burst assembly and constant burst transmission for optical burst switching (OBS) network. In the proposed method, ingress edge node has multiple buffers where IP packets are stored depending on their egress edge nodes, and bursts are assembled at the buffers in round-robin manner. Moreover, bursts are transmitted at fixed intervals with scheduler. To evaluate the performance of the proposed method, we construct a loss model with deterministic and Poisson arrivals, and explicitly derive burst loss probability, burst throughput, and data throughput. In numerical examples, we show the effectiveness of our analysis and compare the performance of the proposed method with Erlang loss system.

Performance analysis of dynamic spectrum handoff scheme with variable bandwidth demand of secondary users for cognitive radio networks

Cognitive radio (CR) has attracted considerable attention as a promising technology for solving the current inefficient use of spectrum. In CR networks, available sub-channels are dynamically assigned to secondary users (SUs). However, when a primary user accesses a primary channel consisting of multiple sub-channels, data transmissions of the SUs already using the sub-channels may be terminated. In this paper, we analyze the performance of dynamic spectrum handoff scheme with channel bonding, in which the number of sub-channels used by an SU are variable. We model the multichannel CR network as a multiserver priority queueing system without waiting facility, deriving the blocking probability, the forced termination probability and the throughput for SUs. In terms of the way of forced termination, we consider two policies; one is that SUs using the largest number of sub-channels are forced to terminate their transmissions, and the other is that SUs using the smallest number of sub-channels are chosen for termination. The analysis is also validated by simulation. Numerical examples show that in both forced-termination policies, the throughput of SUs that are forced to terminate their transmissions degrades as the offered load to the system increases.

Dynamic Hello/Timeout timer adjustment in routing protocols for reducing overhead in MANETs

With the recent advancement and extensive spread of mobile devices, establishing a network with those devices has become more important than ever before. MANETs (Mobile Ad hoc NETworks) have become a promising solution to set up a network with any mobile device. Routing in MANET is still a difficult task since the mobility of the nodes affects the local connections and makes the network topology change constantly. Traditional routing protocols in MANET work sending periodic messages to realize the changes in the topology in order to maintain the local connections up-to-date. However, sending periodic messages at a fixed rate can cause overhead in situations where it can be prevented and thus reduced. In this paper, we introduce a scheme to improve the performance of the routing protocols used in MANET focusing on the reduction of the overhead. In the proposed scheme, instead of sending periodic messages at a fixed rate, we use a link change rate estimation to dynamically adjust the rate at which each node sends the control messages. The link change rate estimation is based on the constant measurement of the local connectivity that reflects the network conditions. We evaluate the performance of our scheme using ns-2 and we compare it to other approaches to show its advantages. With the usage of our scheme, we notice a considerable reduction of the overhead without sacrificing the overall performance of the tested protocols.

A packet scheduling algorithm for max-min fairness in multihop wireless LANs

In this paper, we propose a probabilistic packet scheduling scheme achieving max-min fairness without changing the existing IEEE 802.11 medium access control (MAC) protocol. In the proposed scheme, packets at each wireless node are managed on a per-flow basis. When a wireless node is ready to send a packet, the packet scheduler of the node is likely to select the queue whose number of packets sent in a certain time is the smallest. If the selected queue has no packet, the node defers the transmission by a fixed duration. In order to verify the improvement in per-flow fairness, we evaluate the performance of the proposed scheme by ns-2. The numerical examples show that our proposed scheme achieves better per-flow fairness than the existing schemes in networks of not only chain topologies but also random topologies.

Last-hop preemptive scheme based on the number of hops for optical burst switching networks

We propose a preemptive scheme based on the hop number of a burst for optical burst switching (OBS) networks. In the proposed scheme, when a burst fails in reserving a wavelength at its last hop, the burst can preempt another burst transmission. We consider a tandem network with four OBS nodes and analyze the burst loss probability for the proposed scheme with a continuous-time Markov chain. We also evaluate by simulation the performance of our proposed scheme for a unidirectional ring network. Numerical examples show that our proposed method effectively decreases the loss probability of the burst with a large number of hops.

A preemptive scheme with two-way release message transmission in optical burst switching networks

In this paper, we propose a new preemptive scheme with release message for optical burst switching networks. In the proposed scheme, when a low priority burst is preempted at some intermediate node, two release messages are sent immediately from the intermediate node to both source and destination nodes (two-way transmission), and the release messages release the corresponding wavelengths for the preempted burst. We consider five wavelength selection rules for the preemption and evaluate the performance of the selection rules with simulation. In numerical examples, we show that with our method, the wavelengths are utilized effectively and that our proposed method with the optimal selection rule can decrease the burst loss probability in a large-scale DWDM network.