ts-7000
[Top] [All Lists]

[ts-7000] Re: Flash wear level

To:
Subject: [ts-7000] Re: Flash wear level
From: "marek.zukal" <>
Date: Mon, 09 Feb 2009 12:07:21 -0000
--- In  "Breton M. Saunders"
<> wrote:
>
> Hi Marak,
> 
>   The bottom line is that you don't get any control over the wear 
> leveling when working with an SD card.  The SD cards present themselves 
> as a block device to the system, and allow standard file systems to be 
> used on them - like FAT and ext2/3.
> 
>   In the linux kernel there is support for MTD devices - these are 
> different from block devices in that they expose the OOB (out of band) 
> information on NAND flash chips.  OOB is where the CRC information is 
> kept to verify if a block is valid or not.  MTD devices require a 
> wear-leveling aware file system, like YAFFS2 or JFFS.
> 
>   The "superblock to the filesystem" will be rotated between NAND flash 
> blocks by the sd card's embedded processor, so you're unlikely to see a 
> fault on it any faster than a failure of the entire card.
> 
>   I would suggest that numerous small writes to the sd card are
unlikely 
> to cause problems, unless you intend to deploy your application for a 
> long period (e.g. several years).
> 
>   I would advise you to do accelerated testing - thrash a sd card until 
> it dies using a write pattern from your application - and estimate your 
> application's life time from that.  But remember that your estimates 
> will reflect the specific sd card, and its specific nand flash chip.
 Be 
> sure to put a healthy error estimate in there.
> 
>     -Brett
> 
> > Thank you very much for this answer. I actually thought that it would
> > be that way but I had to try it. 
> >
> > The thing is that I need no huge space for data. Just a few MB but
> > definitely more than any battery backed ram module from TS provides,
> > so the flash is the best way. I am most worried about the superblock
> > of the file system because this is where the most of the stress will
> > be since the files are just small logs. And I have not seen any
> > description of any card that would say "The wear balancing on this
> > card is done by physical block rotation, remapping etc." so it would
> > benefit from the fact that most of the card is actually unused. I
> > found a lot of papers about flash wear balancing but everything was
> > very general. I know that these cards are consumer electronics in the
> > first place but I would expect the manufacturers to be more embedded
> > friendly.
> >
> >
> >
> >
> > ------------------------------------
> >
> > Yahoo! Groups Links
> >
> >
> >
>

Thanks Brett,
I am a little less worried now. We will run some tests and try to
write there as little as possible. There is some tolerance for rare
failures but in general, we expect it to run flawlessly for a few
years 24/7. But if the sd controller is a little intelligent, it
should have a lot of empty blocks to rotate, so it should be fine. 



------------------------------------

Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/ts-7000/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/ts-7000/join
    (Yahoo! ID required)

<*> To change settings via email:
     
    

<*> To unsubscribe from this group, send an email to:
    

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/

<Prev in Thread] Current Thread [Next in Thread>
Admin

Disclaimer: Neither Andrew Taylor nor the University of NSW School of Computer and Engineering take any responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the birding-aus mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU