Zbr's days.
July
Sun Mon Tue Wed Thu Fri Sat
   
7
   
2008
Months
Jul Sep
Oct Nov Dec

About TODO Blog RSS Old blog Projects Gallery Notes

Mon, 07 Jul 2008

New POHMELFS release.

Irish 'Clontarf' and Scotch 'Grant's' helped to rule this release out.

This POHMELFS release features include:

  • Strong cryptography support. One can encrypt whole data channel (except headers) and/or hash/digest it. System will try to autoconfigure itself and if server does not support requested algorithms, mount will either fail (if special mount option is specified) or disable appropriate algorithm usage.
  • Bug fixes.
Cryptography support is essential addition to the POHMELFS core. It was implemented with performance in mind, so that processing speeds would not drop noticeble even in case of very CPU-hungry operations (one can check performance graphs).
POHMELFS utilizes pool of crypto threads (its number can be specified via mount option), which perform data crypto processing and submit it either to network or VFS layer.

Now I will concentrate mostly on userspace server features, mainly its distributed facilities, current ability to write data to multiple servers and balance reading among them is not enough for POHMELFS, but it will be an essential building block of the fully distributed fault-tolerant paralllel filesystem.

If this development will require some changes in kernel side (namely network protocol extension), it will be don in the upcoming releases with possible found bug fixes.

As usual, you can grab sources from archive or via GIT tree.
You can also check POHMELFS homepage to get more details on its design and supported features.

P.S. I think I will have some rest out of this project for several days, which will allow me to concentrate on main POHMELFS features and work out rough edges. I will switch to DST and netchannels (main to make a new releases) and then will devote some time to captcha cracking algorithms.

/devel/fs :: Link / Comments (4)


POHMELFS crypto processing performance.

If you expected a miracle, it did not happen, so I just present a picture, where I compared plain async in-kernel NFS server (no encryption, no checksumming) versus POHMELFS, which performed SHA1 hashing and AES-128-CBC encryption of the whole data channel.
Block size used in iozone test is 8KB, filesize - 8GB, 1GB of RAM.

Encrypted + hashed POHMELFS vs plain NFS

/devel/fs :: Link / Comments (4)