My understanding about lightpeak ( After reading detailed a PC magazine article ) is that it will also be used to connect everything together inside the computer. So CPU, graphics, firewire, usb, ect. will all be connected to this lightpeak pathways.
Correct. So the two buses serve two different purposes. I'm pretty sure that it will not connect Firewire though.
I could but that would require Intel doing that work. I doubt they are in inclined to. The multiplexor is only going to get more
complicated the more protocols stuff on it to handle. PCI-e , ethernet (perhaps) , video , and USB is plenty to handle without
screwing up the dataflow and underlying timing constraints.
LightPeak -- multiplex multiple bus data onto one wire and then demultiplex on the other side. Sort of like how you put video/voice/data onto one TCP/IP wire and then different devices on the ends pull off data they want. There is a required multiplexer/demultplex support chip required on both sides. once to the "other side" the signal is turned back into whatever it was: USB, PCI , video, etc. Additionally, will always be over fiber.
"Light Peak is complementary to existing I/O technologies, as it enables them to run together on a single cable at higher speeds."
http://techresearch.intel.com/articles/None/1813.htm
USB 3.0 -- is over wire. ( at one point Intel wanted fiber but that got rejected in part on expense). It doesn't multiplex anything (except for its own traffic). There are different compatibility mods USB 1.0 , 2.0 , 3.0 but not intended for multiple standards.
Unless running multiple protocol traffic over LightPeak it is doubtful that it will work out as an economical alternative to USB 3.0. The USB 3.0 controllers will be simpler ( once reduce the old legacy 2.0 stuff into new modern 3.0 implementations with process shrinks. ) and cheaper. The wires and connectors will be cheaper. The deployment costs will be cheaper. Every mechanism that LightPeak implementations can use to get cheaper are equally available to USB .
My prediction is it will be backward compatible with usb3 since it is faster technology anyway.
Not. Since Lightpeak must carry more than just USB data will make it incompatible with USB . A future version of lightpeak might transport USB 3.0 in addtion to USB 2.0 traffic ( along with the others defined by the standard) , but that isn't backwards compatible.
I think eventually Mac's will have usb3 connectors but will be connected to lightpeak bus.
Lightpeak connectors are not going to be compatible to USB 3.0
Lighghpeak is going to make for a nice, industry standard "dock connector" . Your computer box is hooked to a separate monitor/Ethernet/usb device all in one connection. Not that it is limited to "dock connector" duties. Apple has repeatedly tried to make the cable connecting the monitor to the computer do multiple things ( display + usb + .... ). Just look at any Cinema display. Likewise the 30 pin iPod/iPad connector. Multiple channels ( USB , device control , etc. ) in one connector.
Lightpeak would be a way to just may that standard.
Lightpeak also allow to chuck the ExpressCard/PCMCIA connectors too.
(or at least do it without tons of moaning. )
That is a different function than USB 3.0. USB is a centralized control of multiple serial connected devices. USB 3.0 just brings enough speed and delivered isochronous capabilites to kill off FW800 and eSATA . This new USB is fast enough to blow away the faster alternatives that have previously existed along side it. That is a different focus than trying to traffic different data into one wire.
USB 3.0 wants everything to switch to "USB only" protocols. That's not Lightpeak's mission. It tries to accommodate multiple protocols transparently to the devices.
It would be a mistake for Apple to blow off USB 3.0. There is a chance though that Apple will attempt to blow away FW to quickly when deploying USB 3.0 though. They have already tried to with USB 2.0, can certainly see it with USB 3.0.
