A systematic method for connection-wise end-to-end delay analysis in asynchronous transfer mode (ATM) networks is proposed. This method consists of the followings: (i) per-stream nodal analysis; (ii) output processes characterization; and (iii) moment matching scheme. Following our previous work, we employ H-MMPPs/Slotted D/1 to model ATM queues. Each virtual connection (VC) in ATM networks can be regarded as a tandem configuration of such queues. In [1], the per-stream analytical results for such an H-MMPPs/Slotted D/1 queue have been provided. In this paper, not only the composite output process is exactly characterized, but also the component in an output process that corresponds to a specific traffic stream is approximated via a decomposition scheme. A moment matching scheme to emulate the per-stream output process as a two-state MMPP is further proposed. Through moment matching, we can then approximate the connection-wise end-to-end delay by recursively performing the nodal performance analysis. The connection-wise end-to-end delay is crucial to network resource decision or control problems such as call admission control (CAC) and routing.

This paper is concerned with spoken language translation in multimedia communication environment. We summarize the current research activities, by describing various spoken language translation systems and the multimodal technology related to spoken language translation. We propose a spoken language translation system that exploits the multimedia communication environment in order to overcome the limits caused by imperfect speech recognition. Our approach is in contrast to that of most conventional speech translation systems that limit their dialogue domains to obtain better speech recognition. We also propose a performance measure for spoken language translation systems. Our measure is defined as the ratio of the information quantities at each end of communication. Using this measure, we show that multimedia enhance spoken language translation systems.