USB 3.0: freakishly fast - maybe

USB 3.0: freakishly fast - maybe

Summary: You'll be hearing a lot about USB 3.0 this year. And well you should, because its potential is vast. But will system vendors step up to the plate to deliver all of USB 3's goodness?

SHARE:
TOPICS: Hardware, Apple
42

You'll be hearing a lot about USB 3.0 this year. And well you should, because its potential is vast. But will system vendors step up to the plate to deliver all of USB 3's goodness?

Speeds und feeds USB 2.0 has never delivered the advertised 480 mbits/sec because that number technically correct and operationally bogus. If you have data transferring in both directions at the same time it could happen - but for USB disks it never does.

That drops the theoretical transfer rate to 240 mbits/sec, but because of protocol overhead - for example, some signal redundancy to increase data integrity - the payload bandwidth is still lower.

Net net: you're lucky to get a 20 MB/sec data rate off a disk - when the advertised rate suggests 60. But unless you use FireWire or eSATA that is the best you can get - until now.

Enter the 3 USB 3.0 is a different protocol - USB is a brand, not a technology - and while I haven't done a deep dive it is a big improvement, while retaining backward compatibility with USB 1 & 2.

The biggest improvement is performance: it can move over 440 MBytes/sec. Here's the 30 second intro video:

The fine print As noted in the video your mileage will vary. We're dependent on the system vendors and their driver writers to develop robust support. That could take years.

Mac users face a bigger problem: it appears that Cupertino is doing nothing - zip, nada - with USB 3.0. With their smaller market share and tighter control, little is likely to happen unless Apple actively supports it.

The Storage Bits take USB 3.0 is a Good Thing. Drives, even flash drives, are getting large enough USB 2 is like sipping the ocean through a straw. The rapid growth of file-based workflows needs more bandwidth - and USB 3.0 looks like a good answer.

Apple is risking their creative professional base if they ignore a fast new I/O bus. Light Peak, an optical interconnect Intel has been working on at Apple's behest, may be their answer.

But as I noted in Light Peak: black hole

Light Peak is a great idea and doomed. Between obnoxious DRM, costly optical hubs and switches, Blu-ray style licensing fees, Intel over-engineering and Apple’s penchant for twee little I/O ports, Light Peak is almost certain to fail.

With Windows 7 momentum and a major I/O fail, Microsoft may be able to take back much of the creative professional market that gives Apple such a hip image.

Let the games begin!

Comments welcome, of course. I don't have any commercial dealings with USB 3 vendors. Saw the demo at the Storage Visions 2010 conference. Update: Sorry about the video - but now it is fixed! End update. Update ][: 3 paragraphs were inadvertently cut and are now restored. Sorry ][!

Topics: Hardware, Apple

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

42 comments
Log in or register to join the discussion
  • You got the spec wrong

    it is not 440Mbps, it is 4.8Gbps http://arstechnica.com/hardware/news/2008/11/usb-3-0-specification-finalized-devices-in-2010.ars

    The Video wont play either, not your issue though, some sort of youtube error.
    Snooki_smoosh_smoosh
    • Who still believes specs?

      Appreciate the note. However I wasn't reporting the spec, but the actual
      number I saw at the demo.

      I've also noted the YouTube error. Hope to get it fixed RSN.

      Cheers,

      Robin
      R Harris
      • Of course specs are always

        theoretical, its just that most of the numbers I
        have seen for this spec were always greater than
        that for eSata @ 3Gbps. And results definitely do
        vary.
        Snooki_smoosh_smoosh
      • RE: USB 3.0: freakishly fast - maybe

        About <a href="http://flvto.com/">convert youtube to mp3</a> once add please, I get choice as of your <a href="http://flvto.com/">mp3 from youtube</a> afterwards to <a href="http://flvto.com/">youtube to mp3 online</a>
        youtube to mp3 converter
    • Don't think so...

      I think you are forgetting bits versus bytes. The spec is 4.8 gigabits per second, the Intel meter is running 480 megabytes per second. That brings the demonstrated transmission rate very close to the theoretical, and with overhead, I would call it a match.

      Robert
      always-a-geek
      • Then confusion came in where in the post

        it is listed as 480Mbps = 480 Megabits per second
        as compared to 480MBps = 480 MegaBytes per second.
        It all has to do with the capitalizations of
        specific letters you know.

        It looks like it may have been corrected since I
        first read it.
        Snooki_smoosh_smoosh
    • MBps vs. Mbps

      The video shows a transfer speed of 446MBps or 446 megabytes per second. The raw interface speed is 4.8Gbps or 4.8 gigabits per second. Multiply the 446 by 8 to convert bytes to bits and you get 3.57 gigabits per second or about 75% of the maximum bit rate. Allowing for overhead, this is a pretty decent result.

      If the original post was edited between you and I reading it, the above may have tormented a few electrons for no good reason.
      DNSB
  • I want PCMCIA cards for USB-3

    So I can retrofix my laptop with USB-3!
    meusterer
    • And likely bottleneck it

      through PCMCIA!
      GoPower
    • With a PCMCIA card you are not going to

      get the speed due to the bottle neck. If your
      laptop has the PC-Express slot then you are
      probably ok, maybe, certainly better than PCMCIA.
      The only thing you would get is compatibility with
      USB 3 devices, however, USB 3 devices should still
      work on a USB 2 port, again resulting bottleneck.
      Snooki_smoosh_smoosh
  • Nobody invested in Video services

    That are using Macs, Final Cut Pro, etc. are goin anywhere soon. Firewire 800 is still great, and if you need faster disk to disk transfers there's always eSATA.
    GoPower
  • RE: USB 3.0: freakishly fast - maybe

    Perhaps Apple already has something better waiting in the wings and wants MS to look foolish. Again.
    tburzio
    • Did you read the article?

      The latter portion of which discussed Apple's approach to faster interconnect.
      de-void-21165590650301806002836337787023
      • NO

        he just wanted an opportunity to slam Microsoft. Because apparently he thinks Microsoft is the driving force behind USB 3. They are not of course because all Microsoft is choosing to do is support it. Light peak sounds cool but what is the cost of the device and the cabling going to be. USB 3 is likely to be cheap and if vendors can support the speed then cheap and fast is good by me. But also the article pointed out that much of the bottleneck in USB2 even is the endpoint devices.
        bobiroc
  • No real reason for this

    eSata should become the chosen interface for external storage, and eliminate the middle-man protocol nightmare. I realize that eSata doesn't carry power, but that would be an easier fix than implementing a whole new protocol.
    terry flores
    • Maybe not

      that would most likely require a whole different pin out to add the pins for power and will ultimately create compatibility problems with older sata devices that do not have the proper pin out. Now if sata was designed with power on the data cable from the start that would be a different story and I suspect that most sata devices require more power than what can be delivered that way.
      bobiroc
      • Not a new problem

        My USB 2.0 Seagate pocket external drive draws more power than is supplied by a single USB port, so it has a special Y cable. And for 3.5" external drives, there is always a power supply involved. So I don't see any huge issue here for 2.5 drives, just add an eSata port and use the existing USB connect for power. Get huge speed increase for the cost of a passthru circuit and an additional port connector soldered to the PCB.
        terry flores
  • DRM on LightPeak? Is that an assumption or part of its spec?

    The last thing anyone wants would be something like an HDCP protocol (or whatever) messing with their I/O devices. Is there any reason (apart from Sony's involvement) for supposing that LightPeak contains such a thing?
    Zogg
    • It's safe to assume LightPeak will need to support DRM ...

      ... because it can be used as a display channel it'll need to support HDCP etc., to ensure the integrity and safety of data from, for example, a BluRay disk.

      It is claimed that LightPeak itself is protocol agnostic and can relay packets of data encoded in any way the transmitter and receiver want.
      de-void-21165590650301806002836337787023
      • LightPeak is protocol agnostic. DRM is dependent on the peripheral. [nt]

        [nt]
        olePigeon