Cloudera

The company was founded in October 2008 in Burlingame ( California ) with a starting capital of $ 5 million, the main goal of the business is the commercialization of the Hadoop project. The founders of the company are Christophe Bishilla ( Eng. Chirstophe Bischiglia ), previously working at Google , Amr Avadalla ( Amr Awadallah , vice president of Yahoo , responsible for analysis and data warehouse), Jeffrey Hammerbacher ( Jeff Hammerbacher , Hive project manager at Facebook ) and Michael Olson , Vice-President of Oracle Corporation, previously CEO of Sleepecat , who developed and developed Berkeley DB and was absorbed by Oracle in 2006) ^[6] . Hammerbacher organized the initial funding of the project by Accel Partners , and Olson led the company. In total, at the initial stage, $ 11 million were attracted, and in addition to Accel among investors, Greylock Partners and business angels Gideon Yu and Caterina Fake are indicated ^[7] .

Hadoop creators Doug Cutting ( eng. Doug Cutting ) and Michael Cafarella ( Mike Cafarella ), former heads of VMware (Diane Green, Diane Green ) and MySQL AB ( Marten Mikos ) ^[8] were among the employees hired in the first months. Due to the fact of Cutting's transfer to Cloudera, the company was characterized as a “new standard-bearer Hadoop” ^[9] .

In 2009, Bishilia was ranked fifth in the list of the 22 best young technological entrepreneurs of the Businessweek weekly ^[10] , and Hammerbacher was placed on the seventh (out of 15) position in 2010 . When nominated by Bishilia, Cloudera was characterized as a service company providing technical advice on Hadoop, while Hammerbacher’s contribution in 2010 was marked as a transformation of the company's business, making it a supplier of replicable software for organizations ^[11] .

In November 2011, the company received additional financing in the amount of $ 40 million ^[12] , in December 2012 - another $ 65 million ^[13] , among investors of the next rounds are Ignition Partners , Greylock , Accel , Meritech Capital Partners and In-Q-Tel ^{[ 14] [13]} .

In October 2012, the company introduced the Impala product., which provides SQL access to data in a cluster running Hadoop, the appearance of such a product was met as a surprise, since the prevailing rhetoric of companies focused on “big data” technologies was the abandonment of traditional technologies based on SQL ( English Old SQL , in consonance with the “ old school ” - old school ) ^[15] .

In June 2013, Tom Reilly was invited to the position of CEO , who previously led two technology companies to be taken over by major players (Trigo, an MDM system manufacturer, was acquired by IBM in 2004, and was launched at IPO and soon absorbed by Hewlett-Packard in 2010), the event is rated as preparation for either initial placement or the sale of a business ^[16] . Olson moved to the post of strategic director and chairman of the board. In July 2013, the firm absorbed the British company Myrryx , founded by Sean Owen, one of the main authors of the scalable machine learning framework which is part of the Hadoop ecosystem , announced the appointment of Owen to the position of “Director of Data Science ” ( eng. director of data science ) ^[17] .

By mid-2013, for five rounds of investment, the company received a total of $ 141 million ^[16] , and in the next round in March 2014, the company raised another $ 160 million ^[18] . In March 2014, after the sixth round of investment, Intel acquired an 18% stake in the company for $ 740 million, thus valuing the Cloudera business at approximately $ 4 billion ^[19] ; at the same time, Intel rejected the development of its own Hadoop distribution a year earlier in favor of promoting solutions from Cloudera ^[18] . In June 2014, the company acquired the Gazzang data encryption technology developer company ^[20] .

In April 2017, the company conducted an initial public offering on the New York Stock Exchange , as a result of which it raised $ 215 million ^[21] . In the fall of 2017, the New York firm-developer of machine learning algorithms Fast Forward Labs was absorbed, the deal was noted as a response to the close integration of Hortonworks with IBM, which focuses on the development of artificial intelligence systems under the Watson program, and abandoned its Hadoop distribution in favor of Hortonworks ^[22] .

In October 2018, a merger with Hortonworks was announced, while the structure retained the name Cloudera, listing on the stock exchange and the general director, and Hortonworks shareholders received 40% of the shares of the combined company ^[23] . The transaction was completed on January 3, 2019, despite the total valuation of the two companies at the time of the announcement of $ 5.2 billion, at its completion the capitalization of the combined business amounted to about $ 3 billion ^[24] . The takeover has actually completed the consolidation stage in the market for commercial Hadoop distributions (of any notable other market participants, only remained with an annual turnover of about $ 175 million for 2018), shifting the focus of competition to wider segments - big data and analytical tools platforms ^[25] .

CDH ( Eng. Cloudera's Distribution including Apache Hadoop ) is an Apache Hadoop distribution, which includes a number of related programs and libraries and Cloudera's own development tools, which are distributed for free and are commercially supported for certain Linux distributions ( Red Hat Enterprise Linux , CentOS , Ubuntu , SuSE SLES , Debian ). Among the Hadoop related Apache software projects, the distribution includes: Flume , HBase , Hive , Mahout , Oozie , Pig , Sqoop , Whirr , Zookeeper . In addition, the distribution kit includes its own cluster management subsystem Cloudera Manager , including Hadoop infrastructure deployment scripts in both local and cloud environments ( Rackspace , Amazon EC2 , ), as well as utilities and configurations to support building automation with Apache Maven .

By the beginning of 2012, two versions of CDH were supplied - CDH2 (based on Hadoop 0.20.1) and CDH3 (based on Hadoop 0.20.2). The CDH3 distribution is included in the delivery of the Oracle Big Data appliance hardware and software complex ^[26] , while the first customer support line for Hadoop is provided by Oracle , and Cloudera provides technical support for more complex issues. In mid-2012, a version of CDH4 based on Hadoop 2.0 (including the YARN module) was released, and CDH4 also includes three of its own products - (Hadoop cluster management browser interface), Impala and Search (full-text and faceted search in HDFS and HBase environments). In 2014 released version CDH5; The CDH6 version, released in the spring of 2018, is based on Hadoop 3.0 (the key innovation of which was the support for noise-tolerant coding for HDFS, which significantly reduces the physical size of clusters) ^[27] .

- a massively parallel mechanism for interactively executing SQL queries on data stored in HDFS and HBase , is distributed under the Apache 2.0 license. Unlike Hive , which translates queries in an SQL-like language (HiveQL) into MapReduce tasks performed in batch mode, Impala performs queries in a distributed environment interactively, distributing the request to processing nodes based on its own mechanism, without resorting to MapReduce.

Cloudera Manager is a specialized component that allows you to automate the creation and modification of Hadoop environments, monitor and analyze the performance of processing tasks, and set up alerts on the occurrence of certain events related to the operation of the distributed processing infrastructure. The annual cost of technical support is about $ 4 thousand per cluster node ^[28] . For Cloudera Manager, there is a free edition ( English free edition ) that works only on clusters consisting of less than 50 nodes and is devoid of a number of properties available to commercial subscribers (such as performance monitoring, configuration version management, Kerberos support ).

Following the Garnter forecast in the data management technologies HYIP cycle of 2017, suggesting that the concept of “Hadoop distribution” itself is outdated in the near future, the company shifted the emphasis in the product offer to thematic kits made up of virtually the same components that are collected in CDH, but aimed at certain specific tasks. So, in 2018, products appeared under the names Data Warehouse (assembly for data warehouses , with a focus on Impala), Operational DB (for operational databases, around HBase , and Spark ), Data Engineering (for ETL and interactive access to data), Data Science (for “ data science ” tasks), Enterprise Data Hub (for enterprise-level data platforms, it is actually a complete assembly of the Hadoop distribution plus a data catalog based on its own SDX component).

The value policy from 2018 is formed around thematic products; Depending on the configuration, subscribers annually pay from $ 4 thousand for supporting each node of Data Engineering and Data Science products to $ 10 thousand for an node of an Enterprise Data Hub product.

• Vance, Ashlee . Hadoop, a Free Software Program, Finds Uses Beyond Search (eng.) (HTML), N. Y .: The New York Times (17 March 2009), S. B3. The appeal date is December 13, 2011.
• Prickett Morgan, Timothy Cloudera floats commercial Hadoop distro (eng.) . The Register (16 March 2009). The appeal date is December 13, 2011. Archived March 11, 2012.
• Taft, Darryl New Cloudera Desktop GUI Simplifies Hadoop for Users (Eng.) . eWeek (2 October 2009). The appeal date is December 13, 2011. Archived May 17, 2012.
• Malik, Om Is Hadoop Champion Cloudera the Next Red Hat? (eng.) GigaOm (2 October 2009). The appeal date is December 13, 2011. Archived May 17, 2012.
• Nairn, Geoff . Big Data, Big Blue and Going Green (English) (HTML), L .: Financial Times (27 September 2010). The appeal date is May 29, 2011.
• Cloudera's Olson Interview About Data Use (English) . Cloudera's Olson Interview About Data Use . Bloomberg (March 22, 2011). The appeal date is December 13, 2011. Archived May 17, 2012.
• 7. Cloudera. Entrepreneur: Jeff Hammerbacher, 27 (English) . Best Young Tech Entrepreneurs 2010 . Bloomberg Businessweek (April 20, 2010). The appeal date is December 27, 2011. Archived May 17, 2012.
• Vance, Ashlee . This Tech Bubble Is Different (English) (HTML), Businessweek , N. Y .: Bloomberg (14 April 2011). The appeal date is May 29, 2011.
• Jackson, Joab . SGI launches Cloudera Hadoop BI clusters (English) (HTML), Framingham: Computerworld (17 October 2011). The appeal date is May 29, 2011.
• Rao, Leena Cloudera Updates Hadoop Management App With Health Checks, Reporting Features And More (Eng.) . TechCrunch (December 8, 2011). The appeal date is December 27, 2011. Archived May 17, 2012.