JindoFS: native cloud computing storage of large data separation scheme
Before JindoFS
Before JindoFS, cloud customers are mainly using HDFS and OSS / S3 as a large data storage. Hadoop HDFS is the primary storage system for 10 years, has become the storage standard HDFS large ecological data, but we can also see HDFS although continuous optimization, but also JVM has failed to break the bottleneck, community and later redesigned OZone. OSS / S3 object storage as a cloud on behalf of, but also large ecological data were fit, but because the characteristics of the object store design, metadata related operations HDFS can not achieve the same efficiency; Object Storage customers increased bandwidth, but it is limited, some of the time difficult to fully meet the needs of users on the use of big data.
Jindo basis derived from
EMR Jindo Ali cloud-based Apache Spark / Apache Hadoop in the cloud custom distributed computing and storage engine. Ji