Abstract:

Dynamic description of Aliyun e-MapReduce

Aliyun e-MapReduce version 1.3 is released

  • E-mapreduce offers an annual, monthly package (about 60% cheaper than on-demand)
  • You can customize software installation and configuration through boot operations
  • In addition, Hbase clusters can be created
  • Open API is used to create clusters and submit jobs

Information broadcast

  • The total investment and financing in the field of big data exceeded 2.4 billion yuan

    The disclosed amount of financing in the big data industry is about 2.42 billion YUAN; A total of 37 investment and financing events occurred in the big data industry; In each subdivision of big data, application financing accounted for 81.1%Copy the code

    In the last nine months, the total amount of investment and financing exceeded 2.4 billion yuan. For example, Tongdun Technology recently announced the completion of 32 million USD B+ round of financing, and the health big data platform Carbon Cloud Intelligence completed A round of 1 billion YUAN of financing estimated at nearly 1 billion DOLLARS. At present, big data is in its golden period and has infinite potential for the future.

  • TensorFlow, Google’s open source AI deep learning system, is going from strength to strength. Machine learning is one of the hottest areas right now, especially after AlphaGo’s 4-1 win over Lee Se-dol. Specifically, TensorFlow uses 100 Gpus to increase the speed of Inception training by 56 times. Spark also supports deep learning and can run on a GPU, such as Yahoo’s Artificial intelligence weapon, CaffeOnSpark.
  • Analysis on how The Warriors become the favorites in the new season: The Golden State Warriors passing network analysis based on Spark GraphFrames was successfully predicted by Ali Yunxiao AI that I was coco Lee to win the championship. Later, some students used Spark GraphFrames to analyze how the Warriors became the favorites in the new season. Predicting future events based on historical information is one of big data’s best skills. The basic reason is that machines can acquire and process more information, and the results become more and more accurate as algorithms are constantly optimized.
  • Continuing to support open source, IBM’s new host platform will better support Spark to help you help yourself. IBM has announced its commitment to the Spark project, deploying more than 3,500 researchers and developers to work on Spark projects in more than a dozen LABS around the world, and providing a ground-breaking machine learning technology — IBM SystemML — for free to the Spark open source ecosystem. IBM will also train more than 1 million Spark data scientists and data engineers. IBM is using Spark to quickly serve its customers.
  • What position is big data using to shape the value of big data? We dig and analyze users’ behavior habits and preferences from behind the massive and complicated data, find out products and services that are more in line with users’ “taste”, and adjust and optimize ourselves according to users’ needs
  • Apache Kylin has some practice in many large companies, and can handle some multidimensional query problems, including Baidu. Big data startup Kyligence, a startup founded by Apache Kylin’s core development team, has announced a multi-million dollar angel round of investment. Apache Kylin, Apache Spark, Apache Kafka, Apache Zeppelin, Apache HBase, Elsticsearch, H2O and others were selected for InfoWorld’s best Open Source Big Data Tool award 2015. Apache Kylin is implemented based on Apache Hbase.
  • Arun Murthy, co-founder of Hortonworks, said HDFS and YARN are very mature as operating systems for big data, and the future is the era of middleware and upper-layer applications. In the computing engine Spark, Flink, Kudu, Sass and other vigorous development stage. In its 10th year, Hadoop has become the standard in the big data industry.