The Design And Implementation Of Big Data Platform Monitor System Of Ctrip Ticket Division

The second revolution of the Internet ushered in the mobile Internet era,resulting in large user demand,big data,high risk of high change,high market-oriented and fast iteration.So how to attract and retain users and dig into the user value,find the user needs,set out reasonable guidance,timely identify risks to avoid risks,clarify the product operation situation and have clear mind of market from compare and analysis to stand out in the fierce market competition are important topics for now.Big data and data oriented product to respond to changes in market are good ways to solve this problem.In the new situation,we have to use big data and new high-concurrence distributed architecture to replace the original traditional application architecture,to deal with the business needs,the data explosion and high iteration.Ctrip big data team make great effort,and have great improved and amazing enhancements by using big data related technology.The main work of the team includes architecture,development and maintenance of big data platform of Ctrip Ticket R&D.Include Hadoop ecosphere,Alibaba open source Hadoop operating platform Zeus,ElasticSearch cluster and ELK platform,Kafka cluster MQ platform,calculation Engine Spark and Storm,distributed query engine Presto,also include the unified monitor platform of Ctrip Ticket R&D.This paper introduces the background and significance of the big data platform and monitor system first.This paper will introduce the business application scenario,the demand of big data and the background of architecture in Ctrip,will also show the main work of big data basic platform group,the group I belong,which include kafka cluster MQ platform,ELK(ElasticSeach,LogStash,Kibana)platform,related data processing technology and unified monitor system of Ctrip Ticket R&D.The related technology involve Kafka,ELK,Guice,Esper,Shiro,Highcharts and so on.This paper describes the architecture and shows the implementation details of the big data platform,and explains the architectural design principles and the big data processing flow by using business examples.The paper will analyze the requirements of the monitoring system,the overall structure of the monitoring system,and the details of each module.The system platform mainly includes FlightDataAggregator(data processing)module and FlightLiveWatch(front end monitoring-display)module.It will combine with the UML diagram and the key part of the code to describe the specific implementation of the monitoring system.It willhave a detailed description of some of the complex or important parts of the page and a partial representation of the page.