Is USB1/2/3 Dead Yet? Long Live Thunderbolt!

I have mentioned before (can't find the thread now...) that Intel's Light Peak will make USB3 obsolete before it even gets off the ground.

That day is today:

See:
Intel Launches Light Peak Tech as ''Thunderbolt''


First to be featured in the new Mac's:

See:
New MacBook Pros Gain Thunderbolt, Goes AMD


Oh, and by the way, 'Light Peak' is now officially called 'Thunderbolt'.

My next external enclosures will be sure to feature eSATA and Thunderbolt connectivity and not USB2/3.

 

Then I'm assuming Intel will never integrated USB 3 into the chipset..

 

Nah, I think it's still on track to be in the Ivy Bridge chipset.

 

but now that light peak is already on sandy bridge.. i'm not so sure anymore intel doesn't like usb3

 

As much as i would love to see Light Peak (i hate the new name for it) take off and be a viable sucessor to everything. I really dread the idea of multipule componet connections flooding hte market vieing for domaniance.

Its like USB3 vs SATA3. Sure SATA3 is a faster connection, but we dont really have much available that can even max out USB3's throughput so wouldnt it be great to have everything just use that simple connection (speaking of USB3) so we never have to worry about anything. Heck no more SATA cables inside of a tower or anything it would be plug and play heaven.

Now if Light peak takes over the market and EVERYTHING goes to that then awesome, but I see another connector war coming and i really hate those

 

No matter how great Thunderbolt is, considering it's only on Apples now and unless something changes radically for there to be widespread adoption, Thunderbolt won't take off - just like FW.

 

the difference: it has full support from intel. which means, it could get widespread.

to kamin: well, light peak can replace usb AND sata at the same time. as it allows both usb and sata protocols to go trough (even at the same time). so while yes, there might be a long time till all other connectors died out, it will rather soon replace stuff internally. so your usb ports of tomorrow will actually be lightpeak behind, as will your esata, and displayport. they're just there so you don't need adapters, they already ARE adapters to lightpeak.

 

Thats what i'm hoping for is a really good unified standard. I really dont care what it ends up being so long as it can be a 1 stop solution. I really hope it is Light Peak because it has amazing through put and has the most chance for up scaling through later revisions.

I just hate that it came out first on Apple, and i think that was the reason Firewire has never caught on. People/Companies see it is for apple and never try the migration over to PC's. Maybe it wont happen this time though and PC's will adopt the format as well. With Intel backing it there is a good chance it will migrate over luckily

 

I too don't really like the name Thunderbolt - 'Light Peak' is much more classy!

As to the Macs having this tech first - don't forget; they'll be behind everything else that is released for the next 3 or 4 years.

Thunderbolt is poised to finally consolidate all the different ways to connect different components as directly as possible to the cpu - I hope that Intel doesn't miss this opportunity to use its sizeable influence to ensure that that does happen.

Unless someone knows of another techology that is inherently better/cheaper/faster/more scalable than LP promises to become?

 

Wouldn't mind if it would take off ... it's kinda an all in one connection.
Usb+sata+pci-e+displayport+ .. in either optical or copper connection ..

Only time will tell.

 

What? Apple is normally company that pushes through the latest technology first, and killing off older connectors first, barring FW as they invented it. That and Apple refreshes their entire line every 6-7 months...

 

I'll agree that they may introduce 'some' new tech first - but a refresh every 6/7 months is no upgrade to me (I've compared a 3 yr old MacBook Pro to current models and find no reason to upgrade (if I even owned one...).

Some might see this as a positive thing (sure, in one way it is...) but what I expect is that the same level of performance will go down in price - but they don't - this is where Apple fails (for me).

 

Another example in addition to FireWire is Mini-DP - pretty much no other laptop outside of the Apple universe uses it. It would be good to see a more unified standard, as "Thunderbolt" seems to be quite a nice successor.

 

Well it's all relative. People are always needing new notebooks, and if they can wait 1 month and the newest models come out, then all the power to them. Many PC manufacturers typically wait at least a year or two before a refresh, or a whole new model.

The refresh last spring was huge, jumping from Core 2 architecture to Core i, though overall design stayed the same.

Apple prices generally stay the same, but now the entry MBP 13" offers Core i series for the same price.

I guess we will see which PC manufacturer standardizes LP first in their model line..

 

All it takes is one pc company to start equipping units with TB (it's a dumb name yes, but get used to it :/ ) and they will all have it. Macs are still the redheaded kid with their firewire and mini-dp might-as-well-be-proprietary connections, but if say, sony or HP or asus started rolling out models with TB, it will become widespread.

The only thing that makes me think it won't flop in the face of USB3 is because it is backed by intel. Otherwise it sounds a lot like e-sata.

 

Its not just about notebook makers...its about peripherals...eSata never got off because there was no external HDD, Camera or other stuff supporting it.
USB3 has some stuff already available...LP does not!

 

eSata certainly is out there, but it is less popular due to the fact that it is _only_ for external drives. Pretty much all my systems have the port.

 

I don't really see thunderbolt taking off as a well used port massively outside of Apple. There's just too many things that would have to be switched round. It would mean adaptors for my keyboard, mouse, speakers, printer, scanner and huey, I'll skip the external HDD as that's on esata. It's not like I'm getting rid of some monstrosity of a port like the old printer serial ports used to be.

So I guess that leaves the question, how much do I need the extra bandwidth? Personally I'd say not as much as I even need HDMI. Correct me if I'm wrong but I believe I'm right in saying there's not quite enough bandwidth for high def video AND audio as there is with HDMI in the current copper iteration of Thunderbolt.

It's coming and I think with Intel peddling yet another new connection we'll all have one of these ports eventually, the real question is who makes use of it. I can see it going the way of my never used firewire port on my old socket 939 PC.

Poor old Intel, I can see this going the way of Meego and mobile phone processors....

 

Considering they've already said some heavyweights in the peripheral industry hopped on (WD in particular), it really does fall on the computer manufacturers to start it. If there are no ports available on your computer you won't buy a peripheral.

Also I know I compared it to esata, but esata is different mostly because it requires external power, not because it's for hdd's only. TB does not - it draws power from the chipset like USB.

 

You need to check who invented USB before you make a statement like that.....

 

I call that cheating ... 2 chip and pci-e x4 speeds.
The wire in middle 10Gb/s .. that is more then pci-e v3.0 (8Gb/s) ...

Will be able to connect anything full speed external ... quite fast external gpu connection over it too ..

If it's implemented will mean the end of every ports not just usb (s-ata/network/hdmi/dispaly/expresscard/firewire/...)

 

When we can add anything to it from Graphic cards, HDD's to even CPU's it would be good news. Of course being Intel it would never allow a CPU to be connected to it.

 

Ask lightpeak to call me when I can buy a portable hard drive, copy some stuff to it from my new home computer, then plug it into my 2 year old office computer so I can copy stuff to that.

Oh, and I need to use that portable hard drive to copy something from my computer to my mother's 4 year old machine.

When I can do that without any extra convertor cables, then USB will be dead.

USB is Dead. Long live USB.

 

really? Having a single unified connector that is that fast and it would totally put you off because you would have to buy a Light Peak to USB connector for legacy technology?

Thats a bit odd. It would be like me not liking USB because i cant connect to serial ports with one

 

Exactly!

 

Just make sure that convertor always works. I've had enough headaches with convertors to realize I'm not ready to buy until the technology has gone through a refresh cycle or two.

At this point lightpeak is a new technology. Convertors cables will take some time to be released, then come down in price and also for all the kinks to be worked out. Based on what Intel is saying, we may not even see any real rollout of lightpeak enabled devices until 2012. Is that when the convertors are coming?

I'm looking forward to light peak. If it becomes ubiquitous and replaces every other connection, great! Good riddance to USB.

Until that point, long live USB.

 

Sigh. Yes please just keep USB 3.0. I'd hate to have to get adapters and/or buy new products, and then in a few years find that an older hard drive or other device I need to use is incompatible, blah blah blah. Have Lightpeak offer the same connector as USB and make it completely compatible and I won't care.

Or offer BOTH for about the next ten years and all will be good. Do you know how many desktops still come with serial ports anyhow? Legacy doesn't die fast, it dies hard. Not to mention things like PS/2 ports.

 

Thunderbolt is purely a holding technology IMHO, Intel couldn't get Light Peak to work using optical in time so Thunderbolt uses copper......expect to see a new iteration in 12 months (or maybe less) probably called Lightning (or some other clever wordplay on Light Peak/Thunderbolt imagery) that does use optical and increases the bandwidth accordingly.

I think that's why the manufacturers aren't going to adopt this yet, USB3 is a clear path forward...why confuse consumers with a new technology that may be replaced in 12 months.

 

Yeah, I was also thinking about this, they couldn't get optical now but you figure they will get it to work most likely. In fact I thought I read somewhere that this technology would start off as copper then optical will follow. Then what? Will the port that uses copper become useless?

 

Yeah, I was also thinking about this, they couldn't get optical to work now but you figure they will get it to work most likely. In fact I thought I read somewhere that this technology would start off as copper then optical will follow. Then what? Will the port that uses copper become useless?

 

Look at what Itanium was expected to do. Time will tell.

 

Do note that Thunderbolt/LightPeak is a wide open, bi-directional data channel with NO embedded security. As it stands now, a compromised external device can easily grab the entire contents of your hard drive without the user knowing about it. This has already been demonstrated and cautioned against by Intel.

Way similar to the initial, and long-standing security problems with Firewire. Security problems with Firewire have really never been taken care of.

In it's first incarnation, all Apple did was to trade one insecure high-speed interconnect for a faster, even less secure interconnect.

Think Different. Or not.

At least USB has a master-slave relationship setup so that at a hardware level external devices just can't grab root/admin access as an uncontested part of their handshaking.

I wouldn't touch Thunderbolt/LightPeak until MacOS and the various Microsoft OSii offer device-driver specific security for the hookup. Preferably security that uses the embedded security instruction sets of the newer iCore processors as a foundation.

 

So you're saying someone connect an external device to your machine and the user not know about it? I'm not following.

 

I think he means if you were to connect a 1/ 10 Ethernet cable via an adapter to the Thunderbolt port (gah! who thought it would be a good idea to name it that... maybe it was the 5 year old son of a chair person at Intel... we can only hope)

 

I don't really see this taking off. Intel's Light Peak is Sony's Betamax. They're stifling competition just because they don't control the standard. NEC has pretty much been the only USB3 player up until now, and until Intel joins in there's going to be a lack of USB3 solutions. Do you really think that anyone is going to want to pay money to license Intel's Light Peak?

Personally I like the idea of having a one in all connector, but Intel needs to support USB3 as well and then let the free and open market decide. Either way Intel wins (they either get more money for supporting USB3 or they get licensing fees out of LP).

Now if Intel made LP an open source standard, didn't charge, and provided a sample LP IC design and the only 'for pay' service was hardware validation to ensure interoperability (much like USB2 and USB3 does)...then you might have something...

 

I don't know about Intel NOT supporting USB 3.0 especially after reading this. Personally I like USB 3.0 and it's backwards compatibility with USB 2.0 ports... means I can keep using my current devices/connections and upgrade if need-be... I'm not so keen on the idea of the multi-purpose port... ESPECIALLY if i've only got one... how am I going to connect my external monitor, external hard drive and external mouse with adapter if they all use Light Peak... and i've only got one darn plug. Personally, i think Apple just wants the "WOW" factor of just implementing Light Peak... Visit their home page "Breakthrough I/O speeds"... if only the average 5400 rpm external drive could fully saturate USB 2.0...

 

Or could be like Blu Ray, while nothing was clear in the beginning with HD DVDs, the standard prevail over time, i can see that happening with lights peak, i just wished intel would adapted usb3 faster, and let their standard earn their spot in the market out of proving to be worth it in all departments.

 

My own feeling is to agree with Greg in that it will not become the de facto standard.

I haven't really looked that much into it but here's an interesting article from Intel.

http://techresearch.intel.com/spaw2/uploads/files/Intel Light Peak Interesting Facts_0610.pdf

• If all the books in the Library of Congress were digitized, they would amount to over 20 terabytes of data (a 2 with 13zeroes after it). If you used Light Peak technology operating at 10 Gb/s, you could transfer the whole library of congress in less than 35 minutes.
  
• If you had an MP3 player with 64GB of storage, it would only take a minute to fill it up with music using Light Peak at 10Gbps.

Well maybe my math is failing me but assuming some protocol overhead 10Gbs would be ~1GB/s so it seems to me 20TB would actually take over 5 hours to transfer. The 64GB seems about right but what would the actual MP3s storage speed be? I don't think it would be anywhere close to that and i don't see 64GB worth of cache being used. I can not help think that some people might buy the technology not realizing this.

At least it seems to be doing better than Larrabee.

Although NBR users are probably not your everyday user, it would have been interesting to have had a poll for the thread.

 

people don't understand that it's not a new port. it's a unification of ALL your ports. so your laptop might still have (for connectors compatibility) a usb port, a firewire port, a displayport port (yay), a hdmi port, audio ports, lan port. but internally, all of this will be over ONE lane, the Thunderbolt.

and as this one internal connection that is all that's needed for all those external ports, why not expose it to the outside, so you can extend your external ports at random?

having one thunderbolt connector exposed in public means you can have gblan trough an adapter, or esata, or usb3, or a hdmi display, or all of them (like those multicardreaders, but now multiconnectors).

and over time, more and more will directly support thunderbolt as the connectortype, if all goes well for intel.

then we can get a macbook air with just one thunderbolt, and it can be the connector to the dock that has all the legacy ports for screen, keyboard mouse, lan, digicam, etc..

thunderbold WILL replace everything. that doesn't mean you will not continue to have usb ports, lan ports, etc on your laptop for quite a while. it will just all go trough thunderbolt in the back.

it's not a new port. it's a -port. port of ports. yo, dawg, i heard you like ports..

 

The biggest advantage that LP/Thunderbolt could have is the ability to pipe multiple monitor connections over the same output. Desktops and laptops could easily support 2-3 (or more) outputs with only a small handful of LP ports.

Single link DVI is about 4Gbps, so LP could support two in one connector and still have enough spare bandwidth for a few USB connections (maybe a few USB3 connections with the caveat that share 8Gbps of bandwidth). Two connectors, especially if they are optical cables, could easily have cables that are smaller than today's USB cables.

Big advantage. Will that be enough to penetrate the market? Only if Intel can convince enough OEMs to try it and they aren't greedy.

 

Dave, would you please post a link of where your getting your information from. As already stated I haven't looked into <s>light peak</s> Thunderbolt, but from what I've seen it seems to indicate Thunderbolt being a controller that operates off 4 PCIe lanes from the South Bridge (PCH) and just provides an external port connection. The traditional internal Sata / USB connections would I think still interface through the South Bridge as before.

 

thunderbolt is described to be able to communicate any other protocoll trough it. that means you can feed it usb-data, sata data, hdmi data, etc. you just need the right "splitter" at the end with the right plugs so you can plug it in as needed.

this is nothing new, this is the default design goal of lightpeak: replacing ANY interconnection needs with just one. and to do so, it allows not only "it's own protocoll" to go trough, but it's protocoll is a wrapper for multiple others to go trough at the same time.

links? heck no, too tired to google it for you.

 

Then don't do it for me, do it for the other readers of this thread so that they can be better informed.


The pic I provided is from Intel's own site and would seem to indicate your statement of all internal devices as being connected through Thunderbolt as incorrect. Yes, Thunderbolt provides a communications bus but is itself connected to the some of the PCIe lanes of the South Bridge internally and acts as a host to provide an EXTERNAL communications medium. There is no internal ThunderBolt BUS as you suggest and it IS just another type of port.


I'm tired too, but I'll provide some links.

Thunderbolt (interface) - Wikipedia, the free encyclopedia

Peripheral - Wikipedia, the free encyclopedia

 

I'm saying that as the tech stands now, when you connect to an external device that device has instant full root-level access to your machine.

Doesn't have to ask, doesn't have to depend on a clueless user that clicks on something they shouldn't have. Root/admin access is granted by the hardware as soon as you plug in a cable.

It's a design fault compounded by device drivers that have no security layers. Oh, and double-compounded by Apple pushing an immature tech into consumerspace.

 

Well here is an interesting tidbit of information.

According to this article on Macrumors we won't be seeing any Thunderbolt PCIe cards.

I do realize that typically only power users tend to add new technologies to their machines via expansion cards. The mainstream users tend to switch to new technologies when they buy a new machine with that new technology integrated into it. So this bit of information doesn't matter to most people.

But this may be an indicator of how tight Intel is being with deployment and implementation of Thunderbolt. This may have some ramifications as far as how willing manufacturers will be to adopt the tech.

If nothing else, we will be waiting for new machines to replace existing one before Thunderbolt (I keep writing Thunderbird!) becomes mainstream. Intel has already said real adoption into the marketplace won't start happening until 2012. It'll then take 2-3 years until we see a lot of people replacing the machines they purchased before and in 2011 for ones with Thunderbolt built in.

So this means we won't really see it catch hold in the mainstream market segment until 2014-2015.

One of the big uses I anticipate for Thunderbolt is in future low power portable devices. However, the chip for it (seen at the bottom of this Anandtech article is pretty big in a physical sense. It won't fit into phones, and probably not even tablets. The tech will really need to be shrunk down, which will take some time.

A lot can happen between today and 2014-2015. I mean, the end of the world is in 2012 right?

I'm really hoping that Lightpeak (please call version 2.0 this) catches hold and takes off. But I've seen a lot of good tech vanish...

Thundercats HO!

 

I dont think so.

Intel's Codename Light Peak Launches as Thunderbolt - AnandTech :: Your Source for Hardware Analysis and News

 

Apple "pushing immature technology", eh? Heheh, figures...