[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: OT: SSD for servers

On 200802290116, Michelle Konzack wrote:
> Am 2008-02-27 21:10:03, schrieb Anders Breindahl:
> > I.e., other factors omitted, you could continuously push data onto such
> > a 4GB 10^6 high-endurance drive at 50% levelling at 10MB/s for
> >  
> >   4*10^9/10^7*10^6 * 0.5 / (3600*24*365.2524) = 6.3 years
[ heavy snipping ]
> > Why would this not suffice for server applications? Especially
> > considering the good random access times of SSDs, and the fact that we
> > will use these drives in RAIDs anyway?
> I have killed a new bueyd 4 GByte "SanDisk Ultra II" (27 Euro in Germany)
> in less then one month writing a copy of my ~/Maildir to it and the read
> it ober "courier-imap"...

Those are flash memory cards, and have a very low-level interface --
i2c, IIRC. Those can't be assumed to do wear levelling on their own, and
if you just used your standard filesystem on them, that is sure to waste
the cells fast.

> A normal 2"5 HDD (PATA) was gone in around 10 Month...  so why should a
> SSD be longer live, if it can not beat a HDD physicaly.

See the Wikipedia article; there's pros and cons for both conventional
disks and SSD's. (And apart from the wearing of SSD's, they're all in
SSD's favour, I might add).

When wear levelling gains momentum, Flash-based memory looks very
mature for even server usage.

And, not to forget: Even cheap SSD's compete with high-end harddrives
when it comes to linear transfer rate ([0] says 100MB/s read, 80MB/s
write), and have <1ms access latency. Couple these, and you get
impressive random read/write speeds, which harddrives have a hard time

[0] http://www.samsung.com/global/business/semiconductor/products/flash/Products_FlashSSD.html

Regards, skrewz.

Attachment: signature.asc
Description: Digital signature

Reply to: