Spotify Shifts To HortonWorks’ Hadoop Platform
Music provider Spotify moves its burgeoning Big Data repository of music preferences from one Hadoop player to another
Spotify maintains a large database of recommendations and playlists, and uses it to deliver music to legitimate customers, in a fashion that competes with other easy options such as free piracy, according to Spotify officials. The London-and-Stockholm based firm is shifting from one Hadoop firm, Cloudera, to Hortonworks in October, apparently in the expectation of better support and responsiveness.
Playing a different tune
“Spotify is moving to Hortonworks from a competitive distribution,” Hortonworks’ President Herb Cunitz told TechWeekEurope. “Spotify is a very experienced user. They’ve worked with Hadoop for years. They want to have support and want to get things fixed if they have a problem.” Hortonworks was also promising to get features implemented in the open source Hadoop ecosystem that Spotify needed, he said.
For its part, Spotify said it was impressed by Hortonworks’ efforts to build the Hive data warehouse on top of Hadoop: “The work they have done to improve the Apache Hive data warehouse system aligns well with our needs, as we use Hive extensively for ad-hoc queries and for the analysis of large datasets,” Spotify’s head of infrastructure Wouter de Brie said in a statement.
The implementation will be the largest commercial Hadoop cluster in Europe, running on 700 nodes, Spotify said.
Hadoop is open source, of course, but players compete in the level of support they can offer to the free code, and the level of their commitment to developing the project and tailoring it to their customers’ needs.
Cloudera appears to be the market leader in Hadoop installations at the moment, according to figures from data integration firm Syncsort, with Hortonworks normally reckoned to be running second. However, Hortonworks claims to be the firm that commits most to the open source Hadoop projects, something which Cloudera hotly disputes.
“We have the majority of the committers and wrote the majority of the code,” Cunitz told TechWeekEurope. Earlier in the year, Cloudera chief operating officer Kirk Dunn made a competing claim: “We have 75 percent of the contributors across all the projects. So we dominate,” he told us, reminding us at the same time that Hadoop founder Doug Cutting founded Cloudera.
Do you know about open source? Try our quiz!