Research Into A Cloud Computing System For Speech Interaction

Cloud computing is a hot topic in recent years, it changes the way that users gethardware and software resources, and it gives companies a new mode to provide thehardware, software and storage services. From the concept of cloud computingproposed for the first time to now, just a few years, cloud computing obtained a fastdevelopment, it almost penetrated into various fields of information andtelecommunications industry, and plays a vital role in some areas. Now, cloudcomputing technology integrated with technologies in other fields has become a newtread.Voice is the most effective and most direct way of human communication, itplays an irreplaceable role in the lives of people. Voice technology plays an importantrole in intelligent human-computer interaction. With the development of science andtechnology and the progress of the society, intelligent speech technology will seepinto all aspects of the human’s life. If voice interactive technology has been widelyapplied in all walks of life, it means that a voice interactive system may respond to therequest of millions or even tens of millions of times per second. In response to such ahuge number of requests in a short period of time, the traditional server mode isdifficult to do, even if able to do it, the price is also very high. Cloud computing wasborn to deal with the huge amounts of data and massive computing, so cloudcomputing combined with voice interactive systems is logical. So how to build acloud computing system for voice interaction has a high value in research and use,which is main research content of this article.First of all, this article studied and analyzed the two open source cloud computingplatform framework: Sector/Sphere and Hadoop, and implemented a cloud computing platform based on Sector/Sphere. Then this article studied the voice interactive systemimplemented in traditional server model, this work paved the way for theimplementation of the cloud computing architecture voice interactive system. Afterthat，Used the toolkit HTK and HTS deployed the environment for speech synthesisand speech recognition, and after configuration and programming, this article realizedthe cloud computing architecture voice interactive system. Finally，the experimentscompared the performance between the voice interactive system implemented intraditional server model and the voice interactive system implemented in cloudcomputing model, and Discusses the scalability of the voice interactive systemimplemented in cloud computing model. The experiments proved that compared withthe voice interactive system implemented in traditional server model, the voiceinteractive system implemented in cloud computing model has more advantages,better prepared to face the challenges of massive user requests, and has a certaindegree of Scalability.