digg

Facebook Connect Join Digg About
See the original image at linux-mag.com

NILFS: A File System to Make SSD Scream

linux-mag.com — It’s difficult to write storage articles at this time and not focus on the upcoming 2.6.30 kernel. Why? This kernel is loaded with a number of new file systems — some of which we’ve already covered, like ext4 and btrfs. Another of the hot new file systems that is in 2.6.30 is NILFS. This file system is definitely one that you should be testing.

Who dugg this? Made popular Jun 4, 2009

Digg DudeWhat is Digg?

Digg is a place for people to discover and share content from anywhere on the web. Digg surfaces the best stuff as submitted and voted on by the community. Take the Tour to learn more.

57 Comments

show profanity settings
expand all
  • TheBadAndy, on 06/04/2009, -7/+75Nurse I'd Like to ***** Silly?
  • mephitix, on 06/04/2009, -2/+46Unix interpretation:

    NILF I'd like to "fsck -s"
  • rjinswand, on 06/04/2009, -0/+42Dude, that volume's a NILF!
  • inactive, on 06/04/2009, -2/+42NILFS make my drive hard.
  • mechnoch, on 06/04/2009, -1/+37I know what you mean. I have a solid state drive right now.
  • O8SERVER, on 06/04/2009, -0/+33All I have is a floppy :(
  • dowelly, on 06/04/2009, -1/+33A file search could be called a 'Nilf Hunter'
  • ha3er0, on 06/16/2009, -0/+30Someone thought it was a good idea to submit the second page of the article. I don't know why. Which I only found out during dupe check. So here is the other submission: http://digg.com/linux_unix/NILFS_A_File_System_to_ ...
  • slashdotordigg, on 06/04/2009, -0/+30Aren't SSDs quieter?
  • rolf, on 06/04/2009, -1/+30Whoosh!
  • BenBenMan, on 06/04/2009, -1/+20While this looks promising, I'm looking forward to the release of MILFS, the file system for those who prefer their data to be more mature.
  • LoneWolf01, on 06/04/2009, -2/+18No, I'd Like to ***** Sally.
  • robinthehood, on 06/04/2009, -2/+17Nerds I'd Like to *****
  • mechnoch, on 06/04/2009, -1/+16... Until you format it with NILFS.
  • draculthemad, on 06/04/2009, -1/+12Except the limitation on SSD isnt the total number of writes, but writes per *sector*.

    Most SSD try to avoid this by spreading it out via wear leveling.
    This file system does that automatically, and it sounds like it actually minimizes writes to the same location by avoiding them almost entirely. You'd have to "wrap" all the way around the disk to get back to wear you started before you even hit the second write to the same location.

    Compare this to other fs like *gag* ntfs. It tries to consolidate free space and ends up writing to the first parts of the disk disproportionately. Even ext spreads files out to avoid that.



  • MrTea, on 06/04/2009, -0/+11no
  • longlukey, on 06/04/2009, -3/+11NILFS: Narwhals I'd Like to ***** Senselessly?
  • BugMeNot2, on 06/04/2009, -3/+9They make no sound at all.
  • solid12345, on 06/04/2009, -1/+7Format it with NILFS and you will hear a quiet moaning sound coming from inside your case
  • envirotex, on 06/04/2009, -0/+5"Yeah, ..., um, I'm gonna have to ask you all to come on in on Sunday too."

    Yeah, ..., um, There's TPS reports on page two!
    http://www.linux-mag.com/id/7345/2/
  • Hellothere123, on 06/04/2009, -3/+8lmao! very cleaver
  • Mo0eY, on 06/04/2009, -1/+6Nanna, I Like Fruit Smoothies.
  • bipolarruledout, on 06/04/2009, -0/+5They are pretty silent most of the time but sometimes you can hear the electrons moving around.
  • synergye, on 06/04/2009, -1/+6Cleave Her?
    Damn near killed her!
  • TheInformer, on 06/04/2009, -1/+5I thought it was the MILF filing system when I first looked at the title. Hot swappable drives work best with those.
  • mechnoch, on 06/04/2009, -3/+7Hellooooooo, Nurse!
  • Myztry, on 06/04/2009, -1/+5This sounds like a good redundancy file system system taking advantage of the removal of seek times and the need to erase/write cylindrical tracks of sectors sequentially. Nice to see what will become the dominant storage medium getting implementation to take advantage of it's traits. A few things worry me though.
    1. Ideally you would never write to the same place again (until unavoidable) due to the writes life of nand chips. An evenly distributabed random re-write allocation would be much preferable to a circular, or garbage collected system. Of course then you get the issue of everything containing a reference causing a cascade of writes as they update.
    2. No allowance for raid. That's just silly as raid (or simply parallel access) is where speed can really take off. Ideally a multi nand chip parallel system would be transparently part of the system allowing many chips to respond at once to saturate the communication bus. Perhaps the idea this will be implemented in hardware is why that haven't included the feature. But single chips are cheaper by GB then multiple chips so it can't be relied on.

    /end ass speaking - Been a LONG time since I've written any low level drive access software.
  • palehorse864, on 06/04/2009, -0/+4Ahh crap, NILFS not NOLFS. nevermind.
  • chard, on 06/04/2009, -1/+4I am still trying to figure out why they had to call it NILFS... Then again sex sells so what better way to sell your file system by giving it a acronym that sounds similar to a sexual one.
  • raydeen, on 06/04/2009, -1/+4It sounds good for regular drives but with the constant writing, I'm not so sure it would be good for SSDs due to their Achilles heal of limited writes. Yeah, 100,000 is quite a bit and the jury is out on just how quickly you'd hit the max but it seems that you'd get there a lot faster using this fs. Please correct me if I'm mistaken in my assumption.
  • Sammi84, on 06/04/2009, -1/+4Internet WIN
  • sageerrant, on 06/04/2009, -1/+3Nah, you're just stupid.
  • chrysalis, on 06/04/2009, -0/+2How does it compare to HAMMER?

    HAMMER doesn't describe itself as a log-structured filesystem, but the end result look the same, doesn't it?
  • JQP123, on 06/04/2009, -0/+2"Most SSD try to avoid this by spreading it out via wear leveling."

    Yes, but this is done at the hardware level. Thus, wear leveling is a given --- a common factor that applies to all file systems and thus not really relevant to any comparison.

    Any file system that constantly writes and constantly creates shadow copies is going to wear an SSD much more than a file system that avoids doing these things. So this file system really will make your SSD scream (as in pain) much quicker than some alternatives.
  • wpyh, on 06/04/2009, -0/+1Page 2 is filled with interesting articles.

    (on osnews...)
  • bipolarruledout, on 06/04/2009, -0/+1NTFS is an excellent file system....for magnetic disks. MFFS is going to be available for flash devices.
  • inactive, on 06/05/2009, -0/+1Bitching rims has a new meaning?

    (spectacle rims)
  • PakoBedejo, on 06/04/2009, -1/+2You just sent my mind to someplace I never wanted it to go...

    Oh god...I wish I didn't know what a Narwhal was!
  • BugMeNot2, on 06/04/2009, -1/+1Vroom!
  • wpyh, on 06/04/2009, -3/+3Another version, another FS....
  • MN1962, on 06/04/2009, -1/+1I so thought of this before I read it which means nothing really.
  • hackosaurus, on 06/04/2009, -3/+3So close to "MILFs"
  • PolarZoe, on 06/04/2009, -4/+4Already having load problems:
    http://rorr.im
  • Jpatano, on 06/04/2009, -3/+3dammit... i totally thought this article had something to do with MILF's
  • yournamehere, on 06/04/2009, -1/+1I heard they're horney all the time.
  • mechnoch, on 06/04/2009, -2/+2Negative indicates lament for sarcasm.
  • Fetching more discussions...
    Show 51 - 62 of 62 discussions

  • Add a Comment

    Login or register to comment!
    Or, sign-in super fast with your Facebook account ...


© Digg Inc. 2009 — Content created and posted by Digg users is dedicated to the public domain | Terms of Use Updated | Privacy Policy
DIGG, DIGG IT, DUGG, DIGG THIS, Digg graphics, logos, designs, page headers, button icons, scripts, and other service names are the trademarks of Digg Inc.