Hi, WebHDFS is faster/scalable as it lets the client directly access a DN to read off of, and hence requires that the client be able to access all nodes of the HDFS cluster. The HttpFs (what hdfsproxy is now 'called') is instead useful as a gateway service if clients are not allowed to access all DN nodes of a given cluster.
On Wed, Sep 19, 2012 at 7:18 PM, Visioner Sadak <[email protected]> wrote: > Hello experts could you judge whether webhdfs is fast or hdfsproxy is fast, > is hdfs proxy slower coz it uses https only or can we use http also in > hdfsproxy, its also mentioned in this below link,could you guys throw more > light on advantages and disadvantages of both techniques...... > > http://hadoop.apache.org/docs/hdfs/r0.21.0/hdfsproxy.html#Tomcat-based+Installation+and+Configuration > in disadvantages section, > > > > Thanks.... -- Harsh J
