by Anton Shilov
09/23/2010 | 11:10 AM
The Light Peak interconnection technology developed by Intel Corp. and widely discussed by the industry these days may be years away, at least from personal computers made by large companies. The reason for that are pretty simple and logical: there is no open specification for Light Peak, the high-speed optical interconnection in its first incarnation will not offer a lot of valuable benefits for the end-user.
"I think there will be some who will use Light Peak, but not the volume OEMs like the Acers, HPs and Dells. They won't have a need for it," said a PC engineer from one of the top-tier PC makers on condition of anonymity in an interview with EETimes web-site.
Interestingly, but Intel withdrew all of its Light Peak technical classes from its Developer Forum agenda earlier this month. Moreover, twelve months after Intel launched its Light Peak initiative to much fanfare its IDF 2009, no new standards body has been formed to deliver public Light Peak specifications, note analysts from IGI Group. Still, Intel itself said earlier this month that components needed for implementation of Light Peak would be available by the end of 2010.
The first incarnation of Light Peak will support bandwidth of up to 10Gb/s, which is two times higher than that of USB 3.0, but what will not bring any significant gains to all available peripherals, except, perhaps, external graphics cards or portable solid-state drives. The implementation of Light Peak is estimated to cost from $5 to $10, if not higher, despite of Intel's promise of $2 per port.
"You will need higher data rates than 10Gbits/s to make differences in apps like video something end users can really see. Twice the data rate for that cost just doesn’t make sense - it is taking profit margin away," said the engineer.
It is undeniable that several years down the road transfer speeds of higher than 5Gb/s will be needed for external peripherals. However, before that optical technologies will have to get more affordable.
"The thing that excites me most is the move to low-cost optics - that is a big win for the industry. I don't think Light Peak is the answer, but it creates a vehicle to get people to think about low-cost optics. [...] I see over the next five years a transition from copper to optics in server and storage systems and eventually consumer products," said the expert.