| With the rapid development of mobile Internet,the way people have access to information is growing explosively.As the mobile Internet has strict requirements on network traffic and mobile devices.There have been many shortcomings in the traditional way of pulling information.Such as the acquired information is not interesting to customers,wasting network data and time,consuming a lot of resource of servers and client terminals.Meanwhile the push techniques come to being.It can send the correct information to interested users,avoiding frequent data request from users,therefore reducing the pressure on the server.It can also save lots of network data and power of cellphones.Although there are several kinds of push services,such as GCM(Google Cloud Messaging),Foreign services encounter many restrictions in China,and local services are not as stable.So this paper is targeting to design and implement a general push model based on Android OS and to apply it on the third party APP.Before designing the model,I have studied lots of push scheme,and introduce their implementation and push process.To make the model more efficient,I also study some high performance techniques on server like Java NIO,Linux epoll and Android.Then analysis has been done on push algorithm and Weighted Round Robin has been found to meet push demands.However this algorithm has some unnecessary calculation issue during mobile push.Hence I find dynamical Weighted Round Robin.The model is divided into three parts:client App?server App and push server.To achieve push function,a more compact,convenient communication protocol is designed following MQTT protocol.In the chapter of the implementation,the paper has a detailed introduction to how to realize the three parts with key techniques.After the model has been designed,the paper introduces the use of the push model.It elaborated the approach to integrate client SDK and server,at the same time it explained the function?the methods and the theory. |
PDF Full Text Request