JindoFS - Cloud on the large data high-performance data storage scheme Lake

JindoFS background

Computing storage separation is a trend of cloud computing, there are some problems in the conventional mutual integration of computing storage architecture, such as the problem of computing power and storage capacity do not match each other when there is a cluster expansion, the user in some cases only need computing power or storage capacity expansion, the integration of traditional architecture alone can not compute or storage capacity expansion, storage and computing separation can be a good solution to this problem, users only need to worry about the computing power of the entire cluster.

OSS-based computing storage isolated

image

EMR conventional separation scheme is based on computation storage OSS provide compatibility OssFS Hadoop file system, a user can access data on the OSS by OssFS, thus OssFS OSS retain some advantages, such as providing mass storage, low cost, high reliability, At the same time there are some problems such as slow file rename operation, OSS bandwidth constraints, high-frequency data access of OSS consume too much bandwidth. The JindoFS addition to the advantages described above can be retained OssFS, but also to overcome the problems of the above OssFS.

Guess you like

Origin yq.aliyun.com/articles/720312