[X4U] Why does anybody ever get a USB external hard drive? snipped for size

[David Ledger]

>From: "John F. Richardson" <richards at spawar.navy.mil>
>My answer. Price is the main reason.

And availability. Non-Apple shops sell USB only, and if you need a 
drive now rather than wait till next time you're in a town with an 
Apple shop that sells FW drives, you'll end up with USB. That's why I 
have a few USB drives.

>Assumption: an external hard drive is an internal hard drive in a case with
>various ports.

Of some quality. May only be PC desktop quality; may be PC server 
quality if you pay for it. May even be XServe / high-end server 
quality - if you pay a lot.

>I see various drives.
>Drive 1 is obsolete except for dirt cheap drives.

But Drive 1 is what high street stores sell, mainly (at least in the UK).

>Drive 1: USB 2
>Drive 2: USB 2, FW400
>Drive 3: USB 2, FW400, FW800
>Drive 4: USB 2, FW400, FW800, eSata
>So, the box specs indicate eSata > FW800 > FW400 > USB 2 [oops, maybe USB 2
>is close to FW400...I forget]. But assume above is true.
>Q 1: What internal drive mechanism is in these drives?
>If a standard IDE drive mechanism, then I claim that USB 2 is
>just as good as any other. You pay more to have the ports included.
>Before eSata, I suspect that this was true.

FW is always better than USB. Whatever rate the bits come/go down the 
wire, with FW the interface dumps/collects the data directly 
into/from system memory; with USB the CPU has to stop what it's 
doing, save state, copy the next byte (or 2 or 4 or 8 ... 4k bytes) 
from/to the interface to/from system memory, restore state, and 
continue what it was doing each time a byte (2, 4, 8, ..., 4k) 
arrives/can be sent. Don't know what the number is, or if it's fixed 
or variable. I've never written a USB driver.

>Definition: native transfer rates = transfer rate from magnetic 
>storage media sectors to the port.

This is in two parts. The transfer rate from medium to some internal 
buffer is always storage bit density x rotational speed. It's the 
transfer from that buffer to the port that can vary. Most drives have 
a several MB buffer that is used as a queue, which complicates 
matters. The drive may also do read-ahead into this buffer, which 
means some data is available immediately and some requires a seek and 
a data transfer. Similarly, some write requests may be satisfied by 
writing to the big buffer, giving 'instant write'; some require the 
data to go to medium before a write is complete. This is on top of 
'buffered i/o' within the system.

>So, if drive 4 really has an internal storage mechanism that has 
>native transfer rates that match the eSata specs, then Drive 4 is 
>the best.

Not all drive boxes with external eSata use Sata drives inside. I've 
seen one review where this was pointed out as a bad mark.

>But what about 1, 2 & 3? Do they have an internal mechanism with 
>native transfer rates in the eSata range? Then it is price. Is the 
>discount worth it?

The board that hangs on the side of the drive is what controls the 
drive buffering and the drive i/o method. You can't get better than 
that whatever you bridge to.

>If 1, 2 and 3 have an internal mechanism with native transfer rates 
>from the storage (seek, latency, cache, etc.,...) to the port 
>(USB2,FW400/800) then YES .. there is no reason to have a USB 2 
>drive except price and you get what you pay for.

Except FW is always better than USB.

>So, is there any way to verify the native transfer rates? I have 
>been assuming that all FW drives just have a bridge chipset that 
>converts IDE mechanism data streams to use USB/FW protocols. I 
>believe that the vast majority of original consumer FW drives, even 
>in 2005 or 2006 had IDE disks inside a case with a bridge chip. I 
>hope I'm wrong.

You'd be right. Before eSata there only was IDE and SCSI in the 
market range. No-one made drive boards with native FW or USB as far 
as I know. They all required a bridge. The market would be too small. 
Most drives end up as internal PC drives, so IDE (and now Sata) is 
fine.

>Please list at least the LaCie native rates for FW drives in 
>2005-2008 time frame. I have asked this (in various forms) several 
>times over the last two years and never received an answer even for 
>just one brand name. What are the drives that actually pump data at 
>FW800 speeds? Any URL's?

There probably isn't an answer to the first bit. It may well change 
from batch to batch as different drives are used. For genuine 
sustained FW800 speeds I'd look at the small drive arrays with 
several drives RAIDed together with RAID5. RAID0 may do it, but I've 
never liked the reduced reliability. No-one seems to do RAID5 over 5 
drives, which is the simplest, most logical solution.

>FW was extremely useful for video transfer from camera / high speed device
>to high performance SCSI or "native FW class" hard disks.

David


-- 
David Ledger - Freelance Unix Sysadmin in the UK.
HP-UX specialist of hpUG technical user group (www.hpug.org.uk)
david.ledger at ivdcs.co.uk
www.ivdcs.co.uk