Even though information in ATM networks is handled as fixed-sized packets (cells), packet-based scheduling is still needed in ATM networks. This letter proposes a packet-based scheduling mechanism that is based on comparison between a packet-based queue and a virtual queue that represents the queue length provided by a cell-based scheduling mechanism. Simulation results showed that this proposed scheduling allocates the bandwidth fairly to each connection.

The explosive increase of traffic in computer communications is a clear sign that we have entered the multimedia information age. To cope with this ever increasing need, economical optical access networks that support burst traffic such as in the Internet are expected to be developed. The ATM-PON is considered to be a promising candidate for such a network, and vigorous efforts in this direction are being promoted worldwide. This paper focuses on accommodating burst traffic in the ATM-PON. In order to do this, a mechanism to transport bandwidth requests from the ONU to the OLT and an algorithm to support dynamic bandwidth allocations based on ONU requests are needed. We have performed a comparative study on bandwidth request methods and bandwidth allocation algorithms, including bandwidth request dependence on time interval and correlation and/or impact between system design parameters. The results of computer simulations are useful in determining how to accommodate burst traffic efficiently in the ATM-PON.

We present techniques to implement fair-sharing on both link bandwidth and buffer space in a switch or router. Together they possess the following merits: 1. solving the counter-overflow problem; 2. avoiding the "credit" accumulation issue; 3. integrating bandwidth allocation with buffer management. The simplicity of this method makes it a viable candidate for implementational use on switches and routers.

This paper presents the system issues involved with the transmission of pre-encoded VBR MPEG video using CBR service. Conventional wisdom suggests that lossless delivery of VBR video using CBR service requires bandwidth to be reserved at the peak rate resulting in low bandwidth utilization. We calculate the minimum rate at which bandwidth must be reserved on a network in order to provide continuous playback of an MPEG encoded video bitstream. Simulation results using the frame size traces from several pre-encoded MPEG bitstreams and several buffer sizes demonstrate that this minimum reservation rateis much lower than the peak rate when a relatively small playback buffer size is used, resulting in much higher bandwidth utilization. Procedures for performing connection setup and lossless realtime video playback between the video server and the client are outlined. Methods for incorporating VCR-like features such as pauseandfast forward/reversefor Video-on-Demand (VoD) applications are presented.

A uniquely-structured Usage Parameter Control (UPC) method named Virtual-shaping is proposed which considers cell arrival time jitter between user and UPC point. The method uses a modified Dangerous Bridge UPC circuit (Sliding window type) and virtually (logically) shapes cell traffic using cell arrival time compensation to offset cell delay variation (CDV). In addition, the proposed method is based on a cell-buffer-less structure and can be realized with reasonable hardware. The method yields precise and accurate monitoring. Computer simulations show that the method offers higher network utilization than the conventional Leaky Bucket based UPC method. The proposed method will make it possible to create more effective B-ISDNs, and more cost-effective broadband VBR services.

The FDDI-II is a high speed and flexible backbone LAN. It can divide the capacity into one packet-switched channel with flexible bandwidth and up to 16 isochronous channels which may be allocated for a variety of real-time services such as video and voice. How to allocate and maintain isochronous bandwidth is an important issue for supporting good services to users. The FDDI-II standard proposed a centralized scheme to achieve this goal. In this paper, we propose a new scheme in a distributed fashion for the management of isochronous bandwidth. Based on our scheme, the network can support various services in a more efficient way.

The effectiveness of traffic shaping for VBR traffic is analyzed. Evaluation results prove that traffic shaping can improve link efficiency for most forms of bursty VBR traffic and that link efficiency gains of more than 250% can be expected without the shaping delay imposing any significant QOS deterioration. Traffic shaping increases the link efficiency to about 80% for traffic with short burst repetition periods. The traffic shaping techniques and analytical results described herein can be employed in the traffic management of future B-ISDN/ATM networks.