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

Re: serial ata



On Thu, 2005-03-24 at 20:02 -0600, bentley rhodes wrote:
> hey i know i probably asked this already.  but, so what!
> sata hard drives run at about (but not necesarily AT) 150mb/ sec.

Er, understand SATA drives run _no_ faster than the same drive mechanics
but with an ATA or UltraSCSI/320 logic.  The SATA signaling allows upto
1.5Gbps, approximately 150MBps theoretical at 10/8 encoding.

Most drives today are typically 40-70MBps burst maximum.  Many ATA, SATA
and even some UltraSCSI drives come off the same line -- depending on
the manufacturer (see my other posts on this).

> i don't see anywhere on tigerdirects web page that say how many devices 
> you can chain though, per controller.  does anyone know?

The _official_ ATA DMA specifications dictate that it is a point-to-
point connection.  That is, 1 end-point controller, 1 end-point device,
per channel.  There is _no_ "addressing" in ATA.  Now legacy EIDE allows
the master/slave setup, which was _never_ designed for ATA DMA.  But
many people use this legacy EIDE compatibility for ATA in violation of
the specs.

With SerialATA, they have finally enforced it.  Absolutely _no_ slave
per port.  Now most SerialATA controllers offer 2-4 ports.  But there is
only 1 device per port, period.  There is a BIOS-ACPI standard called
Advanced Host Controller Interface (AHCI) that allows intelligent,
although host (software) driven storage of up to 32-channels/ports.  But
it's not a hardware controller, just a standard interface for software.

If you want intelligently controlled channels, you want a storage
controller with an on-board ASIC or microcontroller with SRAM and/or
SDRAM.  Then you get queuing and other goodies beyond what even SATA
with NCQ does.


-- 
Bryan J. Smith                                  b.j.smith@ieee.org 
---------------------------------------------------------------- 
Community software is all about choice, choice of technology.
Unfortunately, too many Linux advocates port over the so-called
"choice" from the commercial software world, brand name marketing.
The result is false assumptions, failure to focus on the real
technical similarities, but loyalty to blind vendor alignments.



-
To unsubscribe, send email to majordomo@silug.org with
"unsubscribe silug-discuss" in the body.