(NASDAQ: INTC) INTEL SCREWED UP the best thing about Light Peak, the name, calling it by the marketing name ‘Thunderbolt’. Other than that, the only thing that changed is they lost the optical part, leaving nothing more than an expensive power grab.
Because the branding makes us queasy, we will keep calling it Light Peak, and suggest you do too. Intel isn’t talking any details about the technology, so lets tell you the things they don’t want out about the refreshingly optical free Light Peak.
Light Peak is a bidirectional optical link that moves data across a link at 10Gbps, or twice the speed of USB3. It will pass PCIe and and DisplayPort 1.1, and can be daisy chained for multiple devices. Those are the good sides.
From there things get bad. Intel lost the optical part, so Light Peak becomes Peak. With that, you have to ask if there is a point to bothering with this? The answer is of course yes, it is a power grab to lock the competition out, extract licensing fees from the industry, and cost the users more. It does this very effectively.
USB3 is a much more open standard, and it works quite nicely. Unfortunately it does not make Intel lots of license fees, so they are trying to kill it. If you recall, the first USB3 cable Intel showed at IDF had optical cables embedded in it. That went away after the first showing, and was replaced by the USB3 we see now, sans optical links. The old USB3 became Light Peak.
Since then, Intel has been doing everything in their power to hamstring and hamper USB3’s development, something that is quite easy for them to do since they control 100% of the compatibility testing. They can give the thumbs up or down to anyone for any reason, and reports are that they do. Ever wonder why it took so long to put out USB3 chips, and where the USB3 parts from Intel, the technologies inventor, is? Look no further than the technology formerly known as Light Peak.
Except that they lost the Light. Well, not 100%, there will be an optical extender later this year, think of it as an external optical to electrical converter. Why would you take the main point, and name of a technology, and remove it? For cost reasons. OEMs tell SemiAccurate that Light Peak optical solutions cost about $40 per machine, and that is an outright deal breaker. On top of that, one OEM told us that it basically didn’t work, something that is backed up the the removal of the optical component.
At that BoM cost, it runs about as much as the CPU in a laptop, clearly a deal breaker. If you think that Intel will deliver the optical version of Light Peak for anything that resembles reasonable cost, you are sadly mistaken. The optical version is basically deader than Firewire 800. Instead you get the cheaper electrical version, proprietary yet expensive, the worst of both worlds.
With USB3 comfortably hamstrung, the removal of the optical part of Light Peak isn’t the end of the world. Without chips or support on the majority of PC chipsets out there, you know which technology is going to win, but not on merit.
Did I mention that they lost the light in Light Peak?
Intel has three launch partners, Apple, Sony and Lenovo, but they are only talking about Apple now, anything less may cause a corporate hissy-fit in Cupertino. Once Apple releases their parts, well, now, everyone else will be able to talk about it too, and they undoubtedly will.
Intel is desperately trying to spin this tech as a good thing. It was, but that part was removed. Light Peak is fast, and does work electrically, but it is hard to see it being worth the cost. USB3 is half as fast, far cheaper, and doesn’t lock users in to proprietary and expensive technologies. If this didn’t lock out the competition and force partners to tithe a lot, you would have to wonder why Intel is bothering. Once again, marketing trumped engineering for the no-longer-Light Peak.S|A
Latest posts by Charlie Demerjian (see all)
- You can now buy bare Snapdragon SoCs with 410E and 600E - Sep 28, 2016
- Spin Transfer Technologies talks about their ST-MRAMs - Sep 27, 2016
- ARM adds CMN-600 interconnect and DMC-620 memory controllers - Sep 27, 2016
- Globalfoundries 7nm process isn’t even close to the name - Sep 26, 2016
- ARM upgrades realtime offerings to v8-R and adds Cortex-R52 - Sep 21, 2016