News & Analysis
Comment
Peyrol
"The group is expected to deliver a standard system makers can implement in ...
EDWARDP.SAYRE
There is a clear differentialtion in XAUI Ethernet as well as 10G Ethernet ...
PCI Express takes on Thunderbolt
Rick Merritt
6/22/2011 6:06 PM EDT
SANTA CLARA, Calif. – The PCI Special Interest Group will launch an effort in July to created a cabled version of PCI Express that will take on the Thunderbolt interconnect developed by Intel and Apple. Backers suggest the PCIe approach will be more open and more optimal than Thunderbolt for delivering high throughput I/O to tablets and thin notebooks.
The new cable will be based on PCIe 3.0 which supports up to 8 GTransfers/second. It likely will support a maximum of four parallel lanes for throughput up to 32 Gbits/s and distances no longer than three meters.
While initially focused on copper, the technology is expected to migrate to higher speed copper and optical links. The road map likely leads to a 16 GT/s version based on PCIe Gen4 in about four years as well as an optical version for longer reach and/or higher data rates at some point.
The cable and connector itself are expected to be flatter than those of Thunderbolt. The PCIe cable also will support power to peripherals at levels likely less than 20W.
Details of the new standard will be defined by a working group now being formed. The group is expected to deliver a standard system makers can implement in products before June 2013.
The effort to write the spec could take nine to 18 months. The biggest part of the work is expected to be defining technical requirements and a new connector.
The new spec is aimed at consumer uses for desktop and mobile PCs and tablets as well as their peripherals such as external storage devices. The PCI SIG has a separate cable group, chartered in 2005, that has already delivered a spec for the 2.5 and 5 GT/s versions PCIe 1.1 and 2.0, supporting distances up to eight meters and aimed for use in servers and other data center equipment.
Representatives of the PCI SIG declined to comment in any way on Thunderbolt. However, the initiative is clearly aimed at similar applications including external disk and solid-state drives.
"This will help proliferate PCI Express into new business opportunities," said Al Yanes, president of the PCI SIG, declining to give examples of how it will be used.
"Right now we see a need from our members," Yanes said, declining to comment on Thunderbolt directly.
"There are solutions [like this] in the industry--Thunderbolt is one of them, and some companies are doing own thing," he added.
|
|
|
|
|
LarryM99
6/22/2011 6:44 PM EDT
I like the idea of some competition for Thunderbolt, but I am a little leery of the blurring of the lines between backplane buses and communication channels. Will there be mapped memory on the other end of that three-meter link?What will be the BER for large block transfers? Links like USB 3.0 are built to address the concerns of a comm channel, and you don't expect it to have low latency for direct memory accesses. Am I being old-fashioned here or are these real concerns?
Larry M.
Sign in to Reply
toto555
6/27/2011 1:48 PM EDT
If you have missed this blog:http://www.semiwiki.com/forum/content/390-will-thunderbolt-kill-superspeed-usb.html
you should have a look at it! It clearly explains that TB is just a way for Intel to capture a new chip market, the TB Controller, by using existing technologies (PCI Express and DisplayPort) which are served by other chip makers and their partners IP vendors (for PCIe PHY & Controller, same with DP)...
In other word, ThunderBolt is a way for Intel to kill their competition, make Apple happy... and eventually kill SS USB.
Will it really happen? I guess (and I hope) NOT!
Sign in to Reply
EDWARDP.SAYRE
7/13/2011 2:46 PM EDT
There is a clear differentialtion in XAUI Ethernet as well as 10G Ethernet between cables and backplanes by the use of similar but different standards. The same differentialtion between cables and backplanes can be maintained in PCI-X. The semiconductor technologies in all cases of multi-Gigabit serial are similar. I see no problems.
Ed Sayre, CTO North East Systems Associates, Inc. (NESA)
Sign in to Reply
goafrit
6/22/2011 8:19 PM EDT
I doubt if this will be a competition. Thunderbolt has solidified its position to be worried about PCI Express in this domain. I give it to Thunderbolt
Sign in to Reply
muna_majisa
6/23/2011 12:41 AM EDT
I suspect the success of thundervolt in the market except itel-apple
Sign in to Reply
realistic1
6/23/2011 12:09 PM EDT
PLX has been demonstrating PCIe Gen3 over optical since last year at IDF. You can see a demo at DevCon today.
Sign in to Reply
JimHas
6/23/2011 1:05 PM EDT
TB is nothing more than PCI-e with some routing capability. It makes a lot more sense if the SIG come up with a standard for extending the PCI-e to external devices and evolves from there.
Sign in to Reply
jbreher
6/23/2011 3:16 PM EDT
Drat. Competing standards.
@Larry - I like the idea of cabled PCIe if only to create vendor-interoperable docking stations for notebooks and tablets. I see this development as a huge benefit for end users.
I understand the desire to be free of licensing issues. However, with a 2-year headstart, TB may be fully entrenched by the time the new standard rolls out.
I think the ideal situation would be if this announcement spurs Intel to open source the TB spec. I'm not holding my breath, however...
Sign in to Reply
LarryM99
6/23/2011 9:00 PM EDT
It seems like Bluetooth 3.0/4.0 would be a better choice for a good docking station, particularly for an ARM-based tablet. I still wonder about the tradeoffs of extending PCIe outside of the case. Maybe I'm just being narrow-minded, or maybe I'm unduly biased by a project I was on that tried (unsuccessfully) to significantly extend a VME backplane bus.
Larry M.
Sign in to Reply
docdivakar
6/27/2011 2:19 PM EDT
@jbreher: Cabled PCIe is not just an idea, has been a reality for years, so the evolution to the next generation will be smooth. In comparison, TB has to start from 1.0. Many vendors today make PCIe cables, both copper & fiber (One Stop Systems, for example):
http://www.onestopsystems.com/accessories.php
One Stop also makes PCIe Gen 2 switches that are available today.
The 3m limitation will confine the applications to most top-of-rack applications. As long as there is an uplink from the rack to the next layer of switching, this shouldn't be an issue.
Dr. MP Divakar
Sign in to Reply
yyrkoon
6/24/2011 1:39 AM EDT
This technology, at least in concept has been around for years. I remember reading about the specification well over 3 years ago. At that time, my mind began swimming of the possibilities. It was a great idea I thought.
Fast forward in time, we get Thunderbolt, and now I am thinking "you've GOT to be joking". This time, I hope it gets done right. Assuming there really is a this time.
@Larry M.
Extending PCIe "outside of the case" has already been happening even long before Thunderbolt. AKA the ExpressCard interface found on some( many?) notebooks. Though perhaps maybe not exactly what you're talking about ?
Sign in to Reply
kinnar
6/24/2011 2:11 AM EDT
This is a very good interconnect alternative planned for supporting PCIe on the Tablets and Tabs. But the lengths specified here are some what beyond the need of the targeted systems.
Sign in to Reply
LarryM99
6/24/2011 10:06 AM EDT
The length is my major concern. ExpressCard does give the capability to plug in external cards, but it's generally not used to extend across the room. This isn't a big deal for an interface designed to do that (i.e. USB) but it can be problematic for backplane interfaces.
Larry M.
Sign in to Reply
elctrnx_lyf
6/25/2011 3:36 AM EDT
Multiple standard trying to find their place in both board to board and cable interface is really confusing. USB 3.0 will no more required if PCIe also comes to cable.
Sign in to Reply
elctrnx_lyf
6/25/2011 3:39 AM EDT
If thunderbolt replaces the PCI express inside the chips, then probably the industry will move toward this slowly in the future.
Sign in to Reply
DHarmon
6/28/2011 5:59 PM EDT
Thunderbolt is not intended to eliminate USB 3.0 as an interface.
TB is really only about connectors if you think about it. It muxes existing PCIe and DP signals, sends them out over an active cable, and then the peripheral demuxes the signals and uses either PCIe or DP as dictated by the application. TB does not provide any more BW for the IOs - it uses existing BW.
Now cabled PCIe does offer a higher BW pipe, but as many have pointed out already, it has been around for years and it has done nothing to eliminate USB as an I/O.
Both TB and cabled PCIe offer some intriguing options for docking solutions for small form factor devices such as tablets and NetBooks. Figuring out how to use this for charging the devices is critical as well - and an area where USB is used almost exclusively today.
Neither PCIe nor TB will ever be used for keyboards/mice/etc. directly connected to a device. Therefore USB 3.0 with its backwards compatibility to support these type of low BW devices as well as 5Gb transfers to flash drives makes it ideal for stand alone box IOs.
Sign in to Reply
fdunn
6/28/2011 9:14 PM EDT
TB is Apple pushing Intel for a proprietary interface. As such I only expect to see it on Apple products. It will linger on there but will not catch on in the WinTel domain.
Sign in to Reply
MartinM
6/29/2011 1:37 PM EDT
It is not only Apple using it, also Sony has announced to use it: e.g. see http://www.extremetech.com/computing/88370-sony-vaio-z-thunderbolt-price-macs-look-cheap
Sign in to Reply
Peyrol
6/17/2013 3:25 PM EDT
"The group is expected to deliver a standard system makers can implement in products before June 2013."
So here we are, have they delivered a standard?
Sign in to Reply