If you don't write it, I just might. It would eventually be of use to me, and it would be a fun way to get involved in Hadoop. How I see using this is to set up a small cluster to use as a reliable file store, and use ftp to programmatically access those files. Could be useful for deployment, such as via cfengine for Bcfg2 (that is, store the packages and other binaries in HDFS, let the client retrieve them via ftp).
n
