Design And Implementation Of Email Workflow For Enterprise Email Processing System

Email,as a simple but powerful communication tool on the Internet,is playing an irreplaceable role in everyday business.Many companies possess and maintain their own email systems in order to ensure information security,and use the domain as one of the company labels.Therefore,a sophisticatedly-crafted and well-functioning email processing system is extremely vita for an organization's or a company's office work.This article presented the design and implementation of an enterprise-class email processing system and workflow.Besides spam detection and filtering,this system also enhances other email features by integrating third-party services and extending the current mailing application on top of the processing workflow framework.This system utilizes the reputable open-source spam filtering project,SpamAssassin,as the core module of its spam filtering functionality,but applies feature extraction and neural network algorithms to remedy SpamAssassin's lacking of key word generation rules for Chinese language.This article describes the implementation of this multi-plug-in auxiliary filtering module,including SPF detection,virus detection,blacklisting,and spamtrap detection,etc.It also supports concurrency support,frequency support,as well as characteristic support.The mail processing workflow is implemented based on rule-matching and job queue.According to rules set by users and the principle of rule-matching,the system scans and screens the mail content to determine the next task for execution.The job queue ensures the tasks to be saved and completed,but also increases the tolerant rate of network errors.Last but not least,third-party services are invoked to enhance existing functionalities.By interpreting and analyzing the architecture,this article decouples and reconstructs the system,and implements it based on the asynchronous event-driven network programming model.The testing results indicates that this system has a remarkable concurrent processing capability.Moreover,with this plug-in-like design,this system also supports module extension and uninstallation even while it is running.In the meantime,the addition of data caching and failure recovery measures,substantially improves the scalability,reliability,and maintainability of the system.