Design And Implementation Of Demand Side Platform For Advertising

With the rapid development of advertising computational advertising,which develops with advertising and Internet,has become popular.computational advertising can get the most matching advertisement through a series of calculations.Also it can improve the click through rate(CTR)of advertising,thus increasing the volume of commodities.At present,the mainstream of computing advertising is real-time bidding advertising,while the demand side platform(DSP)is an important prerequisite for advertisers to participate in real-time bidding.Advertisers can customize their needs through DSP to guide user behavior targeted advertising.so as to maximize the interests of advertisers and platforms.The main purpose of this paper is to design and implement a real time bidding advertising management system on the basis of in-depth understanding of advertisers' needs.This system can help advertisers integrate a series of processes of advertising.So that advertisers can easily adjust the strategy according to the advertising effect of the display,without spending a lot of time and energy using multiple accounts to manage their advertising demands.According to the analysis of DSP system,this paper technically selects Spark as a distributed computing framework,Nginx reverse proxy server and Springboot framework for system development work,and uses Elasticsearch to provide retrieval advertising services.Also,this paper divides the DSP into six modules: user management module,account management module,advertising management module,report module,retrieval module and sorting module.The first four modules belong to the advertising part of DSP,advertisers need to use these four modules for user login,account recharging,advertising strategy,advertising material upload and report viewing,etc.;the last two modules belong to the advertising engine part,which does not need the participation of advertisers.After receiving an advertisement exposure request,DSP needs to search and sort through advertising engine,and finally determine an advertisement that can guarantee the revenue,which has the highest similarity with the request and has possibility for users to click.