Dynamic description of Aliyun e-MapReduce
E – graphs team
Version 1.5.0 (in development)
- Add an overall cluster running dashboard
- You do not need to write ids and keys to access OSS, enhancing security
- After a cluster is installed, you can restart the cluster, modify configurations, and install software
1.6.0 version
- Interactive query (Support Hive and Spark)
information
The father of dialogue Hadoop Doug Cutting | large data and open source in the future Main point: the new hardware, the Spark, Hadoop cloud, big data and technology of China, the development of open source
E-mapreduce helps build enterprise-level data warehouse When the business system on Ali Cloud, using E-MapReduce to build data warehouse is also a matter of days
Big Data, why not a simple upgrade of traditional BI? Big data and the traditional BI is the product of social development in different stages, big data for traditional BI, both the inheritance, also have development, from the perspective of “tao”, BI and big data differ in that the former are more inclined to decision making, in fact description is based on the group more generality, help policymakers grasp macroscopic statistical trends, suitable for business operational indicators support problem, Big data is broader, tends to depict individuals, more individual.
Ctrip’s big data practice: High-concurrency application architecture and recommendation system case study users through all kinds of big data, data-driven products is the main means to solve this topic, Ctrip’s big data team has emerged. After several years of efforts, the technology related to big data has brought amazing improvement and help to the business. Taking the user intention service of basic big data as an example, by changing the “one side of thousands” of advertising and column into “one side of thousands”, the conversion rate has been increased several times while improving the user convenience, availability and reducing the cost, which reflects the real value of big data service.
From 0 to 1, the establishment of the Universal recommendation platform of Yihaodian The Precision recommendation department of Yihaodian has gradually built a real-time, highly available and traceable general recommendation platform through continuous exploration. At present, this platform is being used by more and more people in the company. Starting from the background of the emergence of yhd.COM general recommendation platform, this paper explains the overall architecture design, recommendation process visualization system design and recommendation result visualization system design of the platform in detail, and makes a summary in the end. The system can also be quickly built on Aliyun’s E-MapReduce platform.
When it comes to big data, Hadoop and Apache Spark are familiar names. But we tend to take them literally, without thinking deeply about them. Let’s take a look with me at what they have in common.
Hive performance Optimization This blog describes the author’s summary of Hive optimization methods and problems in the work of Hive