15 Comments

Summary:

Last week at the IDF 2009 Conference in San Francisco, Intel unveiled a new next-generation data transfer technology dubbed Light Peak. It’s basically an optical subsystem comprised of lasers, modules and probably the odd Flux Capacitor here and there. The outcome is transfer speeds of up […]

intel logo small

Last week at the IDF 2009 Conference in San Francisco, Intel unveiled a new next-generation data transfer technology dubbed Light Peak. It’s basically an optical subsystem comprised of lasers, modules and probably the odd Flux Capacitor here and there. The outcome is transfer speeds of up to 10Gbps. (By comparison, the upcoming USB 3.0 standard will provide maximum throughput of ‘only’ 3Gbps. So-called “Hi-Speed” USB 2.0, in case you’re not yet impressed enough, manages a measly 480Mbps.)

Furthermore, Intel says the technology can support multiple devices on the same port simultaneously, without the need for adapters or extension dongles, and maintain data parity at cable lengths up to one hundred meters. Wow.

So it’s heart warming to know that Apple originally devised the concept for Light Peak. Engadget reports that Steve Jobs presented the idea to Intel’s Paul Otellini back in 2007. Apple was interested in developing an insanely high-speed interoperable standard capable of shifting huge amounts of data and, “…replace the multitudinous connector types with a single connector (FireWire, USB, Display interface).”

Intel Light Peak Laser On

Speed Demon

The blistering super-speed of Light Peak makes it ideal for driving bandwidth-intensive devices and peripherals on the desktop — say, multiple HD panels — but the real world benefits to ordinary consumers (after all, most end-users do not have high-end desktop hardware) would obviously lie in handheld integration. Because the technology supports multiple devices on the same port, it means a portable device could potentially do-away with the tradition of providing several different I/O ports. Doesn’t this sound typically Apple? Say goodbye to Ethernet, USB, Mini DV, FireWire, optical audio in/out, HDMI and so on. Light Peak offers to replace them all with a single, super high-bandwidth alternative. Fewer ports means fewer components and smaller, thinner devices requiring less power. Presumably, they would be cheaper to manufacture, passing those savings down to the customers in the shape of lower prices. That said, I wouldn’t expect Apple to factor that into its premium pricing philosophy. Apple’s not exactly known for its low, low prices.

So what does it all mean in practical, everyday terms that, you know, matter to ordinary people? Well, right now, it means almost nothing since the technology is still in the developement stage. According to Engadget’s Joshua Topolsky, Apple intends to introduce the new standard (yep, it’s not just a new toy, it’s a whole new standard) in late 2010. Beyond that, 2011 might see low-power variations appear in handheld devices.

The concept behind Light Peak is typically Apple. It offers a real, practical — and aesthetically pleasing — solution to the age old problem of cable-mess while introducing impressive new technology to the industry. Jonathan Ive and Steve Jobs are, it seems, almost pathologically opposed to seeing buttons, ports and other bits and pieces on their devices. In fact, Ive said of the MacBook’s redesign; “We’ve refined and refined every detail in the service of the user, just to get rid of complexity. If something doesn’t need to be there, it’s not there. I don’t know how we could make something any more essential, any simpler…”

And for ultimate minimalism, look no further than the iPhone. It wouldn’t surprise me in the slightest if Ive and Jobs considered a zero-button design at one point in the iPhone’s R&D infancy.

According to Engadget, the roadmap for Light Peak, as it stood in 2007, “…includes Light Peak being introduced to the iPhone / iPod platform to serve as a gateway for multimedia and networking outputs.”

Light Peak would offer an incredible data transfer platform to replace the USB standard used with all iPods and iPhones today. Imagine transferring an HD movie from iTunes to a device in mere seconds, or your entire iTunes music library in about a minute. And there’s no reason why OEMs can’t build Light Peak right into their chipsets and motherboards, allowing for ever-higher bus speeds.

Apple was smart going to Intel with the proposal for Light Peak. Introducing a new toy is relatively easy, but introducing a whole new platform is much harder. By handing the task to Intel, Apple can ensure this new standard makes its way into more hardware than Apple products alone.

USB 3.0 has been a long, long time coming, and many OEM’s are, presumably, invested in long-term licensing deals which allow them to build USB support into their products. At launch, Light Peak would be an expensive new platform (those licenses won’t be cheap!) that initially only Intel’s first-tier partners might support (Engadget tips the hat toward Sony as a possible early-adopter). Most OEMs will likely “wait and see” if the technology proves stable, reliable and economical in the long run. But at least they will consider it. They might not give it any attention at all if it were to appear as an Apple-only technology.

And goodness only knows if we’re about to see the start of another format war, pitted between USB 3.0 and Light Peak 1.0.

One More Thing

Engadget adds that Intel has plans for a new low-power Atom chip, due to launch next year, that will be positioned to compete with chips currently found in mobile devices such as iPhones, netbooks and set top boxes. Says Engadget, “The indication we’ve been given is that that product (coupled with the Light Peak standard) could provide the basis for some “big” MID news in 2010.”

I’m gonna watch the evolution of Light Peak carefully, and I think you should, too. Sure, it doesn’t (yet) promise the wireless goodness of the more mature and proven USB 3.0 standard, which, at the very least, won’t require end users replace all their peripherals with Light Peak-enabled alternatives. But it does offer the kind of wideband, high-speed convergence our increasingly interconnected devices cry-out for.

  1. While I appreciate that Apple may have helped inspire Light Peak, it’s probably a bit of a stretch to say that it was “their idea.” If so, why haven’t they been included in any of the patents? This particular piece of technology has the potential of greatly change how electronic devices function and the potential for royalties is enormous, if Apple really came to Intel with some seriously concrete ideas and design specs (rather than some vague desires as the documents seem to suggest), why didn’t they make sure that they had some sort of claim on the intellectual property? After all, Apple patent’s everything; even UI gestures whose patentability is decidedly suspect.

    Share
  2. Rob, it’s not “a stretch” at all to say that the idea was Apple’s. In fact, you seem to contradict yourself when you say, in one clause, that you appreciate Apple “inspired” Light Peak, but in the next clause, criticise the notion that it was their idea.

    Also, I don’t claim to know who holds the patents to what. You say Apple haven’t been included in “any of the patents” – but you don’t mention or link to any specifics.

    Really, what’s the point in blindly criticising unless you can offer a really justified alternative viewpoint?

    Share
    1. “Rob, it’s not a ‘stretch’ to say that the idea was Apple’s.” Yes, it is. Show me where, outside of the rumors and sources quoted by Engadget, Apple is associated with Light Peak in any way whatsoever? I’ve looked and I can’t find indication of joint patents Apple, statements from either company, or anything on Intel’s website that even mentions Apple’s involvement.

      Compare that to how things worked with OpenCL or Grand Central Dispatch, two technologies Apple was instrumental in developing. Or how they developed Firewire. Their fingerprints are EVERYWHERE on those technologies. Apple doesn’t subtly guide the development of new technology, they want people to know their involvement, it’s good for their bottom line.

      As to your other comments, there is a substantial difference between inspiring an idea (providing feedback about what you would like to see, offering thoughts about how it might be conceived, etc) and actually pulling it off. Your editor doesn’t take credit for your article and the ideas that it encloses, even if she might have had some sort of initial input. Otherwise, you would have to acknowledge her contribution of intellectual property as a co-author. You were the only individual listed on the by-line, which leads me to assume that this article is your intellectual property.

      In another example, you would raise cold hell if I suggested anything other than the fact that Apple invented the iPod/iTunes ecosystem. Yet, other companies were doing the same thing years before Apple rolled out their iconic product line. They “inspired” iTunes, but it would be incorrect to assert that they provided the idea for the store or product line.

      After reading the original Engadget story carefully, I think that here’s what probably happened: Steve Jobs wanted to streamline the connections used for Apple computers and contacted the CEO of Intel to discuss options. There may have even been a bit of back and forth about what designs would look like.

      But your article makes it sounds as though Apple came up with the specs in house and then sub-contracted with Intel to produce the technology. And that, clearly, isn’t accurate. Otherwise, you can damn well be sure that Apple would have it trademarked (as they do OpenCL) and probably patented (like they do everything related to the iPhone). Apple isn’t a research organization, they’re a business. Which means, if they were really so instrumental to creating this tech that they would want their cut in it.

      Share
    2. One other point, I don’t include links to patents or other official documents for one simple reason, there aren’t any. The only first hand evidence for the existence of a partnership is that quoted on Engadget. Everything else, including that Apple “dictated” this technology or that it will be an exclusive to Apple products appears to be pure rumor.

      Share
  3. Apple might have had some input on the spec, but Intel has been talking about optical connections… in public… for at least five years.
    http://www.theinquirer.net/inquirer/news/1047268/intel-speed-of-light-chips–tainted-by-blue-crystal-hue

    … and something sounding like Light Peak almost a year ago:
    http://www.theinquirer.net/inquirer/news/1049915/intel-claims-photonics-breakthrough

    Share
    1. For those who care to track, USB 3.0 is 5 Gbps (not 3 Gb as stated in some articles). Intel’s (now former) VP Pat Gelsinger mentioned higher speed versions of USB using optical technology in his 2007 IDF talks. Other busses like HDMI and DisplayPort already operate at 10 Gbps, mostly over copper, with signs of a growing market for optical versions. Even HDMI LLC announced plans to carry ethernet simultaneously on their cable earlier this year, although at a paltry 100 Mb speed. Optical 10 Gbps has be in use for Ethernet and InfiniBand with optical for some time; in fact Intel had a business unit called Connects Cables build active optical cables, at least until that business unit was sold in entirety to EMCORE of Albuquerque circa Q1’08, supposedly due to it being unprofitable.
      If one is looking at optical technology for any portable application, certain power considerations also need to be addressed properly.

      Share
  4. What agreement would Intel and Apple have for such new technology? Other than Apple using LIght Peak, would intel offer a cut of the profits?

    It’s amazing stuff?

    Share
  5. Regardless of who invented it, the idea of having one cable from which I can connect my computer to my TV, my TV to my DVR, my DVR to a hard drive, a DSLR to to a TV etc. is quite compelling.

    Share
  6. No matter who invented it, this is big news. Screw the naysayers who complain this will cause a single point of failure & only damper the innovations that PC Modders have w/ multiple interfaces for each device.

    This will help companies like Apple & innovators like HP make their systems smaller, faster and with less components. This makes our computers cheaper (in theory) and helps our environment if those cables are recyclable.

    Anyone that’s every built their own PC and scattered USB, Firewire, SATA, Serial, Parallel, IDE, SCSI, Optical, Audio and a video cable around the inside of the case knows that this will change everything. The only problem is, this breakthrough will give modders & PC builders less flexibility. basically, they’ll have to buy a MoBo w/ the interfaces they want and run with it with little room for breathing.

    How long before this arrives? If it’s in the next 24 months then I’m excited! My guess is 2-5 years.

    Share
  7. [...] What is Light Peak – Intel and Apple Fiber connection and protocols The Newton being reborn? Let the Apple [...]

    Share
  8. ” … Furthermore, Intel says the technology can support multiple devices on the same port simultaneously, without the need for adapters or extension dongles, and maintain data parity at cable lengths up to one hundred meters. Wow. …”

    Provided each “simultaneous” device is using a different protocol (packet type & hardware address), otherwise, multiple devices with the same protocols like USB would just create packet collisions.

    Yes, it does appear to be double duplex (four fibers) using multiple, multiplexed (multi-color ?) channels and “built in” ViCel transceivers. Provided the protocol(s) of choice are peer to peer, then “simultaneous” devices might not be a real problem.

    Wondering: If USB (iUSB?) protocol is used, then the upper limit on cable length is almost certainly less than 50 meters because of latency and “firm” handshaking. If FireWire (iFW) protocol is used, then fiber cable could extend to 400 meters …

    The only fly in all this ointment: The Apple / Intel proprietary hold on this “new” technology … someone should start the “foul ball” hue and cry now, or TI, Lucent, NEC, Korean and Chinese ViCel transceiver makers will be cut out of the loop.

    Share
  9. Aside: my friends tell me that Light Peak does not favor any “legacy” hardware protocol, able to deal with USB (iUSB?), FireWire (iFW?), SCSI (iSCSI), EtherNet (iNet ?) … or what have you.

    I suppose there will be some mad rush to the code rooms to “convert” (ala iSCSI) to Light Peak friendly pass through … My advise would be to pick your favorite peer to peer protocol and may the best win …

    Share
  10. Adam J.: ” … The only problem is, this breakthrough will give modders & PC builders less flexibility. basically, they’ll have to buy a MoBo w/ the interfaces they want and run with it with little room for breathing. …”

    Easy: treat the board connector as a miniature PCI slot and find a way to brew an edgeboard connector that fits, then grab your favorite Intel part number (AMD, TI, NEC, Lucent, VIA, … need not apply) … and … and … Well, it may not be so easy after all.

    Share

Comments have been disabled for this post