Improving Amazon S3 Performance - Hortonworks Data CloudImproving Amazon S3 Performance In general, reading and writing data is slower with Amazon S3 than with HDFS, even when virtual clusters are running on Amazon EC2 infrastructure in the same datacenter where the Amazon S3 buckets are located. This is due to the following reasons. On Amazon S3, renaming is a very expensive copy operation. The slow performance of the rename surfaces during the commit
Optimizing Hive Queries
