Research Of The Platform For Social Games And Automatic Retractable Technology Based On AWS Cloud Computing

Currently the SNS(Social Network Site) is one of the world’s most popular internet applications, The register users are more than 1 billion in FaceBook, which is the largest social networking site. In china,the register users of RenRen are almost 100 million yet. The social game is one of the effective ways that keep user stickiness on SNS. With more and more people use smart phone and tablet computer, the MAU (Monthly Active Users) of top 10 social games on FaceBook are more than millions and servers of every game are also more than thousands. Our challenge is how to deal with huge amounts of data and servers cluster, and how to build a social game architecture which has quick response, flexibility and scalability. We do some R&D work as following in this thesis:Firstly, we analyze characteristics of social game architecture and cloud computing development situation in this thesis, Then we also analyze advantages and limitations of all kinds of services cloud computing providing based on social game framework, including IaaS and PaaS. Specially, this thesis focuses on study and comparison of flexibility and auto scaling on cloud computing, and discuses how to choose the right cloud platform to build social gaming framework.We design ESGP (Elastic Social Game Platform) based on AWS hardwares and open source software, which is elstic, flexible, portable and cost-effective. For both periodic and aperiodic traffic load, we implement predictive model of a second moving average time series algorithm with AWS monitoring and feedback system to automatically scale ESGP.Finally, we implement prototype and auto scaling model via Java and Python. We test core features and performance through using Jmeter tool. As a result, ESGP is good for scalability and flexibility by comparing features, performance and cost on cloud computing platform after testing.