Ansari, Abdul Ghani 16MSCS10 Supervisor - Dr. Fareed Ahmed Jokhio creator Department 0f information technology text Nawab shah QUEST 2019 monographic 40p ABSTRACT In this era the use of video is becoming more common now a day on internet. The videos may be related to studies, general knowledge, news, sports, drama, kids' activities etc. User wants to watch good quality videos. There are several video formats available; a user device may support only a subset of these formats. User may be connected with low bandwidth network. Hence, providing a good quality video in real time is not possible for each user. Therefore, video transcoding is used to convert the video according to user requirements. The transcoding process is both compute intensive and data intensive. Therefore, it should be performed at server side in such an environment where computing resources can be provided as per need. Such environment can be a cloud computing environment where transcoding can be performed . The transcoded videos can also be stored for subsequent request. In this thesis different transcoding cost models and storage cost models are used. The transcoded video is stored in video repository as long as it is cost efficient. This cost efficiency is based on popularity score of each transcoded video. A video is popular if it is being accessed buy users. If a particular transcoded video is no more accessed then it is removed from repository to save storage cost. This thesis takes different cost models of amazon cloud for storage as well as for transcoding servers. The results have shown that current storage cost model (US NORTH REGION) is less expensive and better to use it for a transcoding service were both computing and storage resources as required on large scale. R/IMS-19 16-MS(CS)-10