Abstract:

Dynamic description of Aliyun e-MapReduce

  • The e-MapReduce service goes online in zhangbei availability zone.
  • The enhanced E-MapReduce scheduling function supports minute, hour, day, week, and month scheduling.

information

  • Hadoop is not dead, it is the future of big data

Vamsi K. Chemitiganti, managing director of Hortonworks, a Silicon Valley data management company, disagrees with the notion that Hadoop is failing. He wrote a post on his blog about his belief that the Darwinian open source ecosystem is ensuring that Hadoop is a solid and mature technology platform. Heart of Machine has compiled an introduction to this rebuttal, but the contents of this article do not represent the views of Heart of Machine.

  • What is the AI revolution

Ai trading software can make predictions about stocks, bonds, commodities and other financial products by siphoning off huge amounts of data to understand the world. Ai machines can access books, Twitter feeds, news reports, financial data, corporate earnings reports, international monetary policy, even variety show profiles — anything that helps their software understand global trends. Ai can constantly observe this information, constantly learning, constantly optimizing predictions.

  • What does big data tell you about 2017 job-hopping season

The spring recruitment season of 2017 has just passed. In the first recruitment season after the capital winter of 2016, the Internet recruitment market is quietly changing. Perhaps the biggest change is that the bulk of the Internet’s high-end talent is being recruited from small and medium-sized startups to large and medium-sized companies. We collected the interview invitation data of 100 offers of high-end Internet talents in the first quarter of 2017, to show you the “change” and “constant” of Internet talent flow in this spring.

  • Are called AI giants, Google baidu Microsoft Amazon volume difference? In AI, there are a few companies that are often called giants. So between the giants and the giants, which is bigger? Coincidentally, today seems to be an auspicious day for making a fortune. Several Chinese and foreign AI giants, together to hand over the report card of the new quarter.

technology

  • Kafka high availability implementation principles

With its own advantages, Kafka is increasingly favored by Internet enterprises. Vipshop also uses Kafka as one of its internal core message engines. As a commercial level messaging middleware, The importance of message reliability can be imagined. How to ensure accurate transmission of messages? How do I ensure that messages are stored accurately? How do YOU ensure correct consumption of messages? These are all things to consider. This paper starts from Kafka architecture, first understand the basic principle of Kafka, and then through the kakfa storage mechanism, replication principle, synchronization principle, reliability and persistence guarantee step by step on its reliability analysis, finally through benchmark to enhance the recognition of Kafka high reliability.

  • Stripe’s experience with large-scale online data migration

Stripe has hundreds of millions of subscriptions. Doing a large migration with all this data would mean a lot of work. Stripe users expect our apis to be usable and consistent. This meant that we had to be very careful when migrating: the data stored on our system had to have very accurate values, and Stripe’s services had to be available at all times. In this article, Stripe shares how to safely complete a large migration involving hundreds of millions of data.

  • Spark Block storage management analysis

On Apache Spark, the query and storage management of blocks are distinguished by unique Block ids. Therefore, knowing the rules for generating Block ids can help us understand Block queries, how blocks are located in stored procedures, and how to handle mutually exclusive storage/reading of the same Block.