Intel Shows Off First Light Peak Laptop

Barence writes "Intel has provided the first hands-on demonstration of a laptop running its Light Peak technology — an optical interconnect that can transfer data at 10Gbit/sec in both directions — at the company's inaugural European research showcase here in Brussels. Intel has fitted Light Peak into a regular USB cable, with optical fibres running alongside the electrical cabling. Intel provided a visual demonstration of how data is passed through the cable by shining a torch into one end of the cable, with two little dots of light visible to the naked eye at the other end. The demonstration laptop was sending two separate HD video streams to a nearby television screen without any visible lag. The laptop includes a 12mm square chip that converts the optical light into electrical data that the computer understands."

Comments

me first

[Anonymous Coward]

Isn't that called 10Gbit SX ethernet?

Server technology?

[Happy Nuclear Death]

It's nice they've developed a way to transfer data at ridiculous speeds, but it does the average user no good as long as we're using mechanical hard drives. Even a "mere" 1 gigabit network connection outstrips the ability of spinning platters to absorb it. I guess this Light Peak thing is aimed at the server market then?

Re:Server technology?

[Microlith]

It's aimed at reducing the number of different cables on your desktop, I believe.

The initial demo showed an LCD panel, HDD, and at least one other thing running off a single Light Peak chain. Effectively, they want it to replace USB (for data connections), Firewire, eSATA, SATA, SCSI, SAS, DVI, DisplayPort, probably every audio connection you have, Ethernet, and likely more.

Re:Server technology?

[Atmchicago]

Don't you suppose Intel is aware of it, and would like to sell you their SSDs? In a few years nearly all new PCs will sport an SSD.

Re:Server technology?

[sznupi]

I don't think they would demonstrate it using a laptop and sending video to an external screen in case it would be aimed at server market...

But there you have a usage for it - sending video signal; "one connector to rule them all"?

Re:Server technology?

[ryanleary]

It's nice they've developed a way to transfer data at ridiculous speeds, but it does the average user no good as long as we're using mechanical hard drives. Even a "mere" 1 gigabit network connection outstrips the ability of spinning platters to absorb it. I guess this Light Peak thing is aimed at the server market then?

That's not really a fair analysis. HD video is often stored compressed, but needs to be transferred at full resolution uncompressed to the display medium. The DVI spec supports 3.96Gbit/s. HDMI even goes up to 10.2Gbit/s. There are plenty of other examples where a high-bandwidth transport will be useful.

Re:Server technology?

[Mabbo]

The possibilities go beyond just downloading things from the Internet. Intel are hoping to make this the connector for *everything*- your devices, your video, your printer. Frankly, I don't care how or why the do it, but a single cable type for everything, in my eyes, is a dream come true- no more having that box of every different type of wire and connector.

Re:Server technology?

[doogledog]

Ooh and with that unification think of the DRM possibilities!

Re:Server technology?

[Anonymous Coward]

a "mere" 1 gigabit network connection outstrips the ability of spinning platters to absorb it

Yes. Intel is teh stoopid. They don't know the limits of disks. If they did they would not be so stupid and make stupid things like this.

We have the smart people like you to keep us safe from the stupids at intel.

Re:Server technology?

[travisb828]

Someone will develop something that will take advantage of that ridiculous speed, and then someone will develop something that can take advantage of data being transfered at ludicrous speed. Then one day, in the future, computers will go to plaid.

Re:Server technology?

[h4rr4r]

So only one device gets to talk at a time? Sounds great.

Re:Server technology?

[Microlith]

Nice assumption there, I seriously doubt that would ever be the case.

Re:Server technology?

[phantomcircuit]

That's just ridiculous. SSD drives demanded an upgrade to SATA 6.0 Gb/s because they were saturating the SATA 3.0 Gb/s link. Last I checked 3 is bigger than 1.

Re:Server technology?

[Anonymous Coward]

Ooh and with that unification think of the DRM possibilities!

So particular light signals can now be copyrighted? I guess that makes an aldis lamp an infringement device.

Re:Server technology?

[LWATCDR]

Yes and no.
It will probably use Time Division multiplexing.

Re:Server technology?

[jeffmeden]

Ah, er, what's that? SATA2 runs at 3Gb/s because the paltry 1.5Gb/s of SATA1 was outpaced by fast hard drives. This isn't even counting RAID0 controllers that can effectively double that. Now, on to Gigabit ethernet. Even with optimization most find .7Gb/s is the practical limit for things like NFS or SMB. You may do better with dedicated storage systems but you're getting away from consumer-grade technology.

Summary: Is 10Gb/s too much for a modern consumer desktop? No; if you have a lot to transfer you WILL see the difference.

Re:Server technology?

[clone53421]

Ethernet suffers from exactly the same problem.

Oh wait, it doesn’t.

Re:Server technology?

[V!NCENT]

"It's nice they've developed a way to transfer data at ridiculous speeds, but it does the average user no good as long as we're using mechanical hard drives."
What's the problem with most humans? They always seem to want to only advance to the bare miminum required.

How about:
"Yo guys, I got an idea!"
-"Shoot"
"How about making a cable that is so fast that we'll never have to think about the transfer speed anymore?"
-"That'll be awesome!"
???

Re:Server technology?

[alen]

what is the return on investment? these days we have IP KVM's and integrated light out from HP that give you access to a server from a cell phone if need be. how much is this gizmo going to cost compared to existing solutions and how is it going to save our employers money?

Re:Server technology?

[h4rr4r]

Ok so you have 1 "wire" with two optical fibers, both your external cd and your esata drive want to talk to the PC. How do they do that at the same time?

In this case the interlacing is not a huge deal, but there do exist situations where this can be a big deal.

Re:Server technology?

[h4rr4r]

Actually it does. and it indeed causes all kinds of problems if you ever want to use it for very low latency communication. You end up stringing one link for each pair of machines.

Re:Server technology?

[clone53421]

Not seriously in most typically applications, though.

Re:Server technology?

[clone53421]

*typical

Point being, it’s not like this limitation is bound to render it unusable.

Re:Server technology?

[Anonymous Coward]

I have copied large files to one of my hard drives at over 100 MB/s. That's pretty close to a full gigabit (which would be 128 MB/s). I guess you're right that a 7200 rpm drive can't keep up with gigabit, but it's pretty close. I was quite amazed that the copy went that fast. Sure, buffering and such would've been involved, but the result was still that I copied to the drive at those speeds. This is a 1TB WD Black Caviar SATA2 drive connected via eSATA.

It only stands to reason that SSDs will go up from here and far beyond. Might as well have interconnects that will be able to supply data at the speeds emerging devices will be able to handle, no?

Re:Server technology?

[Blazewardog]

You would obviously not put both of those devices in the same chain then.

Re:Server technology?

[Kjella]

Also called "putting all the eggs in one basket"...

Re:Server technology?

[Anonymous Coward]

If numbers can be copyrighted, I'm sure light can too.

Re:Server technology?

[jellomizer]

I think you are misunderstanding how DRM works... All those connections were for digital information you can DRM digital information. You are thinking about DRM over HDMI that is because the previous methods sent analog information to the device.

It is difficult to DRM Analog information (heck lets even call it ARM (Analog Rights Management)). As the main information is easily decoded. Digital Information can be encrypted.

However you must also realize that Analog has a fundamental weakness is that it isn't accurate and cannot be copied exactly. Hence why all the fuss about DRM. Digital Stuff can be copied over and over millions of time and it is still as good as the original. Analog copies after 1 or 2 copies of copies you can tell the difference.

It isn't about the wire or unification of the wire, or the interface it is the software the handles the information the determines DRM

Re:Server technology?

[MozeeToby]

I just put an SSD into my new laptop and I can say with confidence that I will never buy a computer without at least an SSD boot drive. Less than 30 seconds from cold boot to having programs opened on the desktop, many programs open in less than a second. It is a massive performance boost.

Re:Server technology?

[Anonymous Coward]

http://xkcd.com/722/

Think about it, we already copyright patterns of light: movies, programs, games, etc. Patterns = signals.

Re:Server technology?

[pedrop357]

Perhaps multiple wavelengths a la DWDM or something like these 1000BASE-BX10-D and 1000BASE-BX10-U modules from Cisco

Re:Server technology?

[h4rr4r]

I will grant you that, but these high speed interconnects would be way more awesome if they did not have these limitations. Would it really cost a lot more to run 8 fibers per cable so that 4 devices can talk at a time?

Re:Server technology?

[Anonymous Coward]

To convert Mb/sec to MB/sec you divide by 8, correct?

Most SSD drives only burst at about ~250mb/sec, that's burst, not normal speed, sustained transfer rates of an average SSD that claims 250MB/sec seem to be about 120-150MB/sec. Multiply that by 8 and you get 1200Mb/sec, or 1.2Gb/sec. SATA1 is fast enough for this. And when you put drives in RAID, that speed is 1.5Gb per channel, meaning per Hard Drive, so the speed effectively doubles when two hard drives are installed, so even SATA1 can keep up with most SSD drives.

Perhaps in certain situations SATA2 could be required, such as lucky bursts where all the data you need happens to be in the HD's cache, but this is rare, and realistically SATA1 is enough for SSD's.

Re:Server technology?

[Anonymous Coward]

It will time share that channel just like every other bus in a modern computer system (not really that surprising). It will support expressing timing requirements for each data stream. Note it has full bandwidth connections going in both directions.

It is also possible that they will support multiple concurrent data channels using various optical techniques (multiple wavelengths).

Re:Server technology?

[clone53421]

What if you wanted to have 8 devices talking at a time, though? Or 10? Or 16?

A serial bus is scalable right up to the point where the packet collisions make it impractical.

Re:Server technology?

[Coren22]

At least until a competing standard pops up due to the cost of the patent fees.

Re:Server technology?

[Yetihehe]

But just like with usb (it stands for Universal Serial Bus), you will still need several types of cables for various music players and cameras. Of course Apple will have their own standard, Sony too. Also every phone will have their custom connector just because they really need those 20 unused lines. Add to this all those LPUSB converters, LP hubs...

Re:Server technology?

[Anonymous Coward]

HD video is often stored compressed, but needs to be transferred at full resolution uncompressed to the display medium

Not necessarily: http://en.wikipedia.org/wiki/High-bandwidth_Digital_Content_Protection

Re:Server technology?

[psbrogna]

Huh? How is digital more accurate than analog? Isn't an infinite state system more accurate than a two state system?

Re:Server technology?

[Neil Hodges]

Multiplexing or packet switching would be a way to get that to work, since none of those devices alone will saturate the link.

Re:Server technology?

[vux984]

Analog copies after 1 or 2 copies of copies you can tell the difference.

Of course, one would digitize the first copy from analog sans drm, and be able to reproduce it millions of times from there without further degradation.

Re:Server technology?

[Spykk]

You do realize that cable television was an analog signal that required a special box to decode for years...

Re:Server technology?

[lukas84]

No, there are plenty of SSDs that can sustain sequential read speeds above SATA2.

Re:Server technology?

[jmichaelg]

One application is to reduce the cost of fiber to the home. If you can eliminate the $150 transceiver that converts fiber to ethernet, you've lowered the cost of entry into that market. Getting that cost down will help new companies and neighborhood coops make ftth more common.

Admittedly, the biggest cost for fiber to the home is the labor required to string fiber but even that cost is rapidly declining as various technologies have appeared that allow digging holes without tearing up a street have shown up.

Nonetheless, multiply the $150 by the number of households you have to serve to start up a cable operation and you're looking at some serious change.

Re:Server technology?

[Shatrat]

Precision is not accuracy. What are you doing on Slashdot?

Re:Server technology?

[unek2ku]

its true, that unification think of DRM possibilities, i agree with u

Re:Server technology?

[Anonymous Coward]

Vote parent up, the only reason we don't have ARM is because the analog connection is not reliable enough for any secure encryption scheme. Degradation has nothing to do with it. It is sad to see people on slashdot who still believe that a video tape is the ultimate storage device for movies.

Re:Server technology?

[Ungrounded Lightning]

You do realize that cable television was an analog signal that required a special box to decode for years...

Actually it mostly needed the special box just to tune it - frequency shift the normally-structured 6 MHz bandwidth TV signal to a frequency where the TV would receive it.

Early "premium channels" were "encrypted" by inserting a strong but narrow-band interfering signal in an otherwise empty slot in the signal, near the sound carrier. This would intermittently "capture" the FM sound decoder and paint bars across the video, jamming the picture and sound. Subscribers had a narrow-band notch filter installed in their feeds to remove it.

(There were other systems, too, including one used on "air" channels which selectively lowered the strength of the vertical and horizontal sync signals to below the level of the video. A subcarrier in the sound provided the information necessary to identify and boost the sync signals back to normal.)

Re:Server technology?

[Lumpy]

All you need is ONE analog copy, then digitize it and no further degradation.

I do this all the time, I record HD over component analog into a encoder. I encode to mpeg4 and several billion copies of that all look fantastic.

In fact, I find it hard to find someone that can tell my 720p analog copy is a analog copy.

This is why I love the analog hole, it's very useful and bypasses corporate stupidity.

Re:Server technology?

[Lumpy]

I get that with windows 7 and a 10,000 rpm drive.

but I also get a 320gig drive for the same price as your 80gig.

Re:Server technology?

[spazdor]

In this case, "Decode" was a computationally trivial analog signal transformation. It was like an RF version of ROT13.

Re:Server technology?

[Anonymous Coward]

You mean something like porn?

Re:Server technology?

[debrain]

Sir –

Incidentally, Macrovision was the dominant analog signal encryption for quite some time. Macrovision was easier to decode - as you note, albeit difficult to decode perfectly.

Re:Server technology?

[Anonymous Coward]

The server market already has InfiniBand. Why can't they just use that?

Re:Server technology?

[jeffmeden]

You are being quite careless with those 'b's and 'B's, a bit is b and a byte is B, SSDs can do 250MB/s, *not* 250Mb/s... A modern platter HDD can easily do 150 MB/s (reading the innermost tracks) which is 1200 Mb/s. However, most activity benefits from 'burst' transferring cached information from the HDD, which can be as fast as 230MB/s or 1840 Mb/s, easily outstripping the 1200Mb/s you get from SATA-I. So that's why we have SATA-II; put two of those on a RAID0 controller and you can see some ridiculous speed, easily outpacing even the theoretical 1000Mb/s speed of gigabit ethernet.

Re:Server technology?

[socz]

Fiber Optic technology actually is pretty interesting. Light doesn't shoot through the cable like light out of a flashlight. The light bounces like this /\/\/\/\/\/\/\/ inside the cable.


So you could do this: XXXXXXXX

And that's why you can't coil FO cable toooo much because the light won't reflect back into the cable.

So I figure with 2 'channels,' you can get quite a bit of bandwidth. Also you can have the lasers at different frequencies. So you can stack em and rack em!

Re:Server technology?

[karnal]

Because you can typically charge more for 3 iterations to get to the same "awesome speed" as you can by just going to "awesome speed"....

Re:Server technology?

[spazdor]

The same way you manage to have more than 4 TCP connections open at once, despite your Ethernet cable only having 4 pairs of copper.
The idea of multiplexing and timeslicing on a shared medium is already used at pretty much every level of a modern computer system.

If you have a situation where the performance cost is a huge deal, you probably have a situation where more hardware in parallel is warranted.

Remember 15 years ago, when you could put 2 IDE hard drives onto the same bus? Remember how sometimes you'd put one onto the secondary IDE bus instead because that was faster? Same shit, different signalling medium.

Re:Server technology?

[Anonymous Coward]

Ok so you have 1 "wire" with two optical fibers, both your external cd and your esata drive want to talk to the PC. How do they do that at the same time?

In this case the interlacing is not a huge deal, but there do exist situations where this can be a big deal.

Plug them into two ports? You are assuming that the machine only has one port. How many USB ports does your current machine have? Just the one?

Even if you are daisy chaining or running through some kind of "hub" TDM (Time Division Multiplexing) would work as neither of the devices could saturate the link. Perhaps some sort of QoS (Quality of Service) could be implemented to insure that latency sensitive data was given more slices than non-sensitive data...

Gift horse... mouth.

Re:Server technology?

[Khyber]

"However you must also realize that Analog has a fundamental weakness is that it isn't accurate and cannot be copied exactly."

Which makes digital doubly useless as you then introduce further loss trying to replicate an analog signal, preserve it in a digital format, then re-convert it for output from an analog device (no matter how 'digital' your LCD or Plasma screen claims to be) you are analog and thus it must output in analog.

Looks like digital is JUST AS LOSSY, especially when the original source is analog.

Re:Server technology?

[Gr8Apes]

Even a "mere" 1 gigabit network connection outstrips the ability of spinning platters to absorb it.

1Gb/s = roughly 120MB/s, which is only a little more than a WD black 1TB+ drive can handle, provided you can actually get a full Gb/s flowing on your network. If you have 2 of these drives striped... well, the network will definitely be your bottleneck.

Then again, not too many people regularly move 1+GB of data across their networks, so 1Gb is fine.

Re:Server technology?

[Khyber]

I don't even think he's aware that our first broadband cable modems were analog.

Re:Server technology?

[Khyber]

No, they won't, if Intel decides to tightly control the connector specification.

Re:Server technology?

[jgagnon]

With the speed at which this thing operates, you could replace the vast majority of interconnects on a single system with just one cable. If you need precise timings for a given connection then you could easily run that connection through its own cable.

I see no reason why computers won't have more than one of these ports available, just like most computers come with 3+ USB ports (or 6+ on desktops, typically). Worst case is you add a PCI-X card for another port or two.

Re:Server technology?

[Anonymous Coward]

and the battery in your laptop lasts what, 10 minutes?

Re:Server technology?

[drewhk]

By your logic, we would be able to send arbitrary amount of data in arbitrary short time - which is obviously impossible.

Hint: think about the noise on any physical channel

Re:Server technology?

[timeOday]

Seek times from SSDs are hardly on the same chart as any hard drive, almost irrespective of RPM.

Re:Server technology?

[Anonymous Coward]

It is difficult to DRM Analog information (heck lets even call it ARM (Analog Rights Management)). As the main information is easily decoded. Digital Information can be encrypted.

Don't confuse encryption and DRM...they're two different concepts. DRM may use encryption, but it's much more than that. The analog equivalent (analog?) of encryption is scrambling and was done long before anything digital was ever encrypted. Scrambling isn't nearly as effective as encryption, but it does exist.

ARM, as you termed it, is harder because structure around the signal isn't there, not because the signal can't be protected.

Re:Server technology?

[timeOday]

It's nice they've developed a way to transfer data at ridiculous speeds, but it does the average user no good as long as we're using mechanical hard drives.

If we get to where most client computers have super-fast networks, why store any data on them at all? OK, a little bit of your data is private, proprietary, etc... but the vast majority of it (by file size) is just other peoples' IP (games, OS, media, music)... why store it locally?

Re:Server technology?

[spire3661]

You mean like Apple did with Firewire? Good plan.....

Re:Server technology?

[drewhk]

You have only to places of distortion: the ADC and DAC. In fully analog systems every part of the analog chain adds to the degradation of the signal. In modern information processing systems and long communications lines this is too much.

Re:Server technology?

[drewhk]

http://en.wikipedia.org/wiki/Master/slave_(technology)

Re:Server technology?

[obarthelemy]

Speed is mice, but is not the only thing:
- if I look at my PC, I see: 8xUSB2, 4xUSB1, VGA, HDMI, DVI, e-SATA, 6xsound,FireWire, SD/MMC, Ethernet... having just one standard to rule them all would be nice
- apparently, there's also power in that cable. that good, I wouldn't use it without

Extra speed will be nice, in a forward-thinking way, but to me that's not really the main thing. eSATA/USB3/Ethernet speeds are plenty fast enough for me right now.

Re:Server technology?

[Fieryphoenix]

This would be awesome. I have hungered forever for an affordable way to have my noisy machine isolated far away from the area I actually use my PC.

Re:Server technology?

[warrior]

Making a copy of digital information is either lossless or the copy is in error. The making of an analog copy is lossy due to any or all of thermal, electrical, and mechanical noise. The analog copy is also lossy due to bandwidth limits and nonlinearities of the components of the analog system.

Re:Server technology?

[Anonymous Coward]

3: Profit

Re:Server technology?

[Changa_MC]

Reproducibility is not accuracy either.
Records are more accurate than CDs, even though an MP3 from CD is probably better than a tape from a record.

Re:Server technology?

[dgatwood]

From Airplane II quotes:

Soldier: Those lights are blinking out of sequence.
Murdock: Make them blink in sequence.

Re:Server technology?

[V!NCENT]

Intel, in this case, doesn't seem to be doing that. So what's up with the /. comments that it is unnessecairy? (which was my point)

Re:Server technology?

[Anonymous Coward]

http://en.wikipedia.org/wiki/Carrier_sense_multiple_access_with_collision_detection

Re:Server technology?

[MobileTatsu-NJG]

Also called "putting all the eggs in one basket"...

Are you saying life was better back when we had PS2, Serial, Paralell, and SCSI ports on our machines?

I don't mean that to sound snide. I'm actually curious: Do you feel like with all this reliance on USB we've painted ourselves into a corner?

Re:Server technology?

[tdelaney]

Actually, I've managed to achieve just under 900mbps actual throughput using my home network:

1. Intel gigabit NICs at each end (on mobo, no jumbo frames);

2. 3x 1.5TB drives in JBOD (no striping, hence single-drive performance, fully defragged) for reading;

3. G.Skill Falcon II 64GB SSD for writing;

4. Windows 7 (64-bit) at both ends;

5. ASUS GX1108N unmanaged gigabit 8-port switch;

6. Cat-6 cable (approx 20 metres between machines, switch about 2/3 of the way along).

The test was (IIRC) about 25GB of sequential data, and I tested using a stopwatch - I wasn't relying on what Windows was telling me. I only did the measurements because things were moving so much faster than expected (previously with older hardware and Windows XP I'd only been getting about 400mbps max throughput).

In synthetic tests, my hard drives gave 125MB/s (1gbps) sequential read speed, and my SSD gave 114MB/s (0.92gbps) sequential write speed.

So I'd say currently that consumer-level equipment is just about able to saturate gigabit ethernet once you take into account the TCP/IP overheads. It won't be long before consumer-level storage outstrips gigabit.

BTW, now that the system has been up and running for a couple of months, I find I normally get about 700mbps throughput.

Re:Server technology?

[osu-neko]

It's nice they've developed a way to transfer data at ridiculous speeds, but it does the average user no good as long as we're using mechanical hard drives. Even a "mere" 1 gigabit network connection outstrips the ability of spinning platters to absorb it.

A and B are interdependent components of a system that determine its ability to perform task T.

Someone suggests improving A, but skeptics point out that it does no good, because B won't be able to keep up and T will still take just as long.

Someone suggests improving B, but skeptics point out that it does no good, because A won't be able to keep up and T will still take just as long.

Someone intelligent, who wants to see T eventually be done faster and better, applauds either effort when it comes along, and shoots the skeptics because they are stoopid...

Re:Server technology?

[petermgreen]

Yes you DO take some loss of quality in the intitial digitisation and in the case of video also in the compression that is needed to get it to a manageable size and you also take some loss in the reconstruction.

However no matter how many copies of that digital file you produce you do not take any further loss and at least for audio and video the loss of the digitisation, compression and reconstruction compares favourably to even one generation of storage in analogue formats that are available to consumers.

Re:Server technology?

[Bob-o-Matic!]

what is the return on investment? these days we have IP KVM's and integrated light out from HP that give you access to a server from a cell phone if need be. how much is this gizmo going to cost compared to existing solutions and how is it going to save our employers money?

To answer your first question, it seems good for Intel to make the CPU the bottleneck again, that way we finally feel the need to buy new CPUs. As for home use, my MiL's current, C2D 3.06GHz, iMac does not seem appreciably faster than my C2D (Merom) 2.33GHz Macbook Pro which I have had since October of 2006. My informal Handbrake benchmark (crunching video is my only real use for fast computer gear these days) is out of whack due to a recent switch to 64bit Handbrake and 64-bit nightly build of VLC, but they seem to be in the same league, definitely not worlds apart as a 3.5 year difference in tech should have yielded.

Re:Server technology?

[petermgreen]

One application is to reduce the cost of fiber to the home. If you can eliminate the $150 transceiver that converts fiber to ethernet, you've lowered the cost of entry into that market. Getting that cost down will help new companies and neighborhood coops make ftth more common.
I very much doubt a system designed for short distance hookup between desktop devices (and therefore probably based on relatively cheap but lossy optical components) will be compatible with a system designed to send data kilometers.

Nor would I think you would want to connect a local desktop interconnect system directly to a relatively untrusted line.

Re:Server technology?

[drsmithy]

Ok so you have 1 "wire" with two optical fibers, both your external cd and your esata drive want to talk to the PC. How do they do that at the same time?

Same way a computer with only one CPU can run multiple processes at "the same time".

Re:Server technology?

[drsmithy]

Ah, er, what's that? SATA2 runs at 3Gb/s because the paltry 1.5Gb/s of SATA1 was outpaced by fast hard drives.

Rubbish. There isn't a SATA hard disk on the market that exceeds 1.5Gb/s, even in synthetic benchmarks, for anything except a direct-cache transfer (which is irrelevant due to its size). Though they are getting close, with 10k RPM drives and high-density 2TB drives hitting ~140MB/sec. In real-world use it's unlikely most people would notice even if bandwidth was a "paltry" 0.75Gb/s.

This isn't even counting RAID0 controllers that can effectively double that.

No, they're irrelevant because you can only put a single drive on a channel. The RAID controller has multiple channels [0].

Summary: Is 10Gb/s too much for a modern consumer desktop? No; if you have a lot to transfer you WILL see the difference.

No, you won't. Not on anything that could be referred to as "a consumer desktop" with a straight face (hint: multiple RAID arrays with double-digit spindle counts and 10Gb networks aren't "consumer" anything).

[0]Except for port multipliers, but they usually come with their own internal bottlenecks.

Re:Server technology?

[n17ikh]

Replying to this because I accidentally modded you "overrated" instead of "informative". Slashdot's new moderation system needs a confirmation dialog!

Re:Server technology?

[wagnerrp]

Encryption != compression

Even if the video feed is encrypted with HDCP, it is still raw uncompressed video.

Re:Server technology?

[English French Man]

My physics lessons a few years back scream at me that the XXXXXXX "2channels" is not possible. Light does not reflect that precisely through the fiber.

Multiple wavelength works great though.

Re:Server technology?

[English French Man]

6+ ? Last motherboard I bought had fourteen.

Re:Server technology?

[English French Man]

Demonstration on a laptop says: "Look, it doesn't use much power", regardless of what it really uses.

Re:Server technology?

[RMH101]

I'm just glad that I've got digital ears, so I can listen to music without any A/D conversion. No, wait..

Re:Server technology?

[Hurricane78]

So that’s why it was so easy to circumvent that “protection”. Come on. I build notch (and other) filters for my first rudimentary synth when I was still in puberty. It’s piss-easy. You could teach it to any Joe Random in a couple of minutes.

Re:Server technology?

[Hurricane78]

Well, we could simply transport out boolean variables in 64kbit integers, and waste 65535 bits instead of the 63 bits we waste nowadays.

(I’ll Barf a Yogurt with bit field flavoring into their faces when that happens. ;)

Re:Server technology?

[Hurricane78]

What's the problem with most humans? They always seem to want to only advance to the bare miminum required.

Most humans? More like: All life in all of the whole universe!
It’s called efficiency. Look it up! ^^
(You can always use those resources for other things, where they are actually urgently needed.)

Re:Server technology?

[dave420]

Exactly - you have to have at least one lossy conversion, so why insist on more as if it won't hurt the final listening?

Re:Server technology?

[fgouget]

It's nice they've developed a way to transfer data at ridiculous speeds, but it does the average user no good as long as we're using mechanical hard drives.

You do realize that LightPeak is also meant to replace things like HDMI cables, right? HDMI 1.3 has a peak video bandwidth of 8.16Gb/s so that already leaves you with only 1.84Gb/s for your file transfers, something which SSDs will easily fill. Or you could just max it out by connecting two lower resolutions screens.

But LightPeak has advantages other than bandwidth. With it you can connect your computer to the TV on the other side of your living-room, or even on the other side of your house. Try doing that with HDMI or DVI!

Re:Server technology?

[plan10]

Why would one life form tell another life form to be more efficient, when that one life form benefits from the other being inefficient?

That is a very inefficient way to go about things.

Re:Server technology?

[GigaplexNZ]

Records are more accurate than CDs

Only because of the Loudness War. CDs are quite capable of outperforming vinyl in terms of noise, distortion and dynamic range, as long as some music exec doesn't order the mastering engineer to butcher the crap out of it.

Re:Server technology?

[GigaplexNZ]

You could teach it to any Joe Random in a couple of minutes.

You overestimate the abilities of Joe Random. Poor Joe is probably one of those fools who continuously tries to push a door open despite the large "Pull" sign, can't figure out how to set the clock on the VCR/microwave/other device, and is the target audience of the "do not eat" labels on random packing materials.

Re:Server technology?

[Anonymous Coward]

Macrovision was a copy protection system, not encryption.

Re:Server technology?

[Anonymous Coward]

Analog is lossy through every single step. Digital is only necessarily lossy in the initial encoding and the final decoding stages.

Which were you actually arguing for? I couldn't quite tell.

Re:Server technology?

[Khyber]

You do realize this new fiber link is likely being joined with the USB 3.0 + optical spec, right?

It's already under control, and that fiber is going to necessitate a minimum jack size.

Re:Server technology?

[LinuxIsGarbage]

But I like having my printer plugged into my scanner, plugged into my zip drive, plugged into my license dongle, plugged into my PC.

Re:Server technology?

[MobileTatsu-NJG]

Heh. Have you ever used software that used parallel port dongles for copy protection? For those peeps, USB was a gift from heaven.

Re:Server technology?

[psbrogna]

You're making my point for me- digital is more precise than analog (ie. more reproducible), but less accurate than analog (ie. capturing the original information more faithfully).

What are you doing here on Slashdot?

Re:Server technology?

[karnal]

Some posters can't see past their own noses, I'm guessing. Having faster-than-currently-usable interfaces just means that there will be a longer period of time to where the interface will be useful, in my opinion.

Computers are Hard

[Anonymous Coward]

...into electrical data that the computer understands.

Wow, this summary of the article is really technical. Can someone help me understand?

Torch into one end with two little dots of light

[Anonymous Coward]

Intel provided a visual demonstration of how data is passed through the cable, by shining a torch into one end of the cable, with two little dots of light visible to the naked eye at the other end.

The second little dot was a floating-point error.

Finally

[Anonymous Coward]

maybe now I wont have any bottlenecks to my 768k DSL line

or my 12x dvd drive, oh something that can keep up with my sataII hard disk??

the future looks awesome

Optical light?

[jo_ham]

As opposed to... mechanical light?

There's my new patented method for data transfer. Measuring the impact of photons on a force transducer.

Re:Optical light?

[Jamu]

I guess it just means it uses visible light, as opposed to, infrared or ultraviolet for example.

Re:Optical light?

[binarylarry]

A better question is... they used a TORCH to show off the optical cable?

Re:Optical light?

[h4rr4r]

Yes, using a flashlight is a pretty normal way to do that.

I suggest you learn English.

Re:Optical light?

[DarkKnightRadick]

those silly Brits. They call a flashlight a torch.

Re:Optical light?

[Anonymous Coward]

Those silly Yanks, they call a torch a flashlight.

Re:Optical light?

[Coren22]

Yes, using a flashlight is a pretty normal way to do that.

I suggest you learn British.

FTFY :)

Re:Optical light?

[Anonymous Coward]

When the computer was invented, a new word was coined for that too.

Actually, the word "computer" has been in use for hundreds of years. Before the advent of mechanical and electronic computers, it referred to a person whose job was to perform computations.

Re:Optical light?

[mweather]

I've never once heard someone refer to a burning stick as a flashlight.

Re:Optical light?

[Anonymous Coward]

Context my friend, context. Surely too advanced a concept for you uneducated yanks. A torch can just as easily mean a burning stick in the UK, for that is what said item is called. But in the context of stuck in the wild, make a torch, it's painfully obvious that one would be looking to make a burning stick.

Without context, most of language does not work. Or are there really *no* dual-meaning words in the US English???

Re:Optical light?

[nevillethedevil]

You weren't at our last office party.

Re:Optical light?

[Grishnakh]

Sure, there's dual-meaning words, but I'll bet most of them we inherited.

If there's something new, why reuse an old word and rely on context, when you can just easily create a new word instead and eliminate confusion? It works well for the Germans, after all, and they're much more famous for their engineering prowess and technology than the Brits are, so they're obviously doing something right.

Re:Optical light?

[Quantumstate]

"When the cellphone was invented, a new word was coined for that too. Why aren't the Brits trying to call it a "telephone" or "smoke signal" or "megaphone" instead?"

You might be interested to know that in the UK people call them mobiles which is short for mobile telephone, some people also use the word phone to refer to a mobile phone as well.

Also from what the Oxford English Dictionary says:

Computer:

1. A person who makes calculations or computations; a calculator, a reckoner; spec. a person employed to make calculations in an observatory, in surveying, etc. Now chiefly hist.

The current usage is actually a conversion of the word because a new word wasn't coined when we switched from people to machines doing the computing.

Your other three examples seem to be correct. However your case is weakened dramatically. I will add a couple more counterexamples to show how common it is to reuse an old word.

Bus has been in continuous usage from early horse drawn models to modern combustion engine (or even electric) designs.
Screen has been changed from a thin flat piece of paper/wood to a computer screen.

Re:Optical light?

[TheRaven64]

Sure, there's dual-meaning words, but I'll bet most of them we inherited.

Check? In written English, there is no ambiguity between a cheque and a check, in US English, there is, and you also use check to mean bill, which adds another layer of ambiguity (you use a check to pay the check after you check it, we use a cheque to pay the bill after we check it). There are a number of similar cases where homophones have different spellings in English but the same spelling in US English.

Re:Optical light?

[tolydude]

Torch is the standard British term for flashlight.

Re:Optical light?

[Grishnakh]

You might be interested to know that in the UK people call them mobiles which is short for mobile telephone, some people also use the word phone to refer to a mobile phone as well.

To me, "mobile" is an adjective, not normally a noun. There is a noun form of the word, but it's pronounced mo-BILE, and it's a thing that hangs over an infant's crib. (It's also the name of a city in southern Alabama.)

Screen has been changed from a thin flat piece of paper/wood to a computer screen.

I don't know many people that call monitors "screens" any more. They're normally called "monitors".

If you say "screen" to me, I'm immediately going to think of the mesh things in my house's windows that keep bugs from flying in when I open the windows for fresh air. I also think of metal screens that many houses here in Arizona have over the windows to keep the sun out (they're made of aluminum, with the cuts in them angled so that sunlight from overhead is blocked). I might also think of silkscreens, which aren't very different from window screens. As for paper/wood, I've never heard of a screen made from those materials. Screens have been mainly fabric for hundreds of years.

I'll grant you the bus one, but even there it doesn't sound like it's describing something entirely different, just a vehicle meant to transport a larger group of people. The only thing that's changed is the form of locomotion. Of course, I could say the same thing about carriages (now automobiles). But then again, most people call them "cars" now, and I believe that word far predates the automobile.

Re:Optical light?

[Grishnakh]

In written English, there is no ambiguity between a cheque and a check,

Yes, but in spoken UK English there is.

American English has had a trend (even since America declared independence I believe) of changing spellings to make them simpler and resemble the spoken form more. The "-ize" suffix is an example of this, as well as dropping the superfluous 'u' from words like color, honor, etc. Any word with a "que" on the end instead of "ck" would have gotten the same treatment. "Cheque", after all, looks more like a French word than an English one (and may very well be one for all I know).

Re:Optical light?

[MarkRose]

You've obviously never made a torch out of a branch off a magnesium tree!

Re:Optical light?

[jo_ham]

This is why it was such an annoyance that the US version of Harry Potter was not only renamed "sorcerer's stone" (wtf, Americans won't know what a philosopher's stone is?), but references like "torch" being changed to "flashlight" just... dilute it.

It's no wonder people have trouble with terms they don't hear often.

Re:Optical light?

[Anonymous Coward]

woosh?

Re:Optical light?

[Anonymous Coward]

This is why it was such an annoyance that the US version of Harry Potter was not only renamed "sorcerer's stone" (wtf, Americans won't know what a philosopher's stone is?), but references like "torch" being changed to "flashlight" just... dilute it.

It's no wonder people have trouble with terms they don't hear often.

Reminds me of an interview with Neil Gaiman I read. Apparently his American publisher just did a simple find and replace to remove Britishisms from one of his books. The result was that the proofs came back with phrases like "...he said apartmently."

Re:Optical light?

[Anonymous Coward]

I am pretty sure Grishnakh is a brit trolling as a stereotypical arrogant dumb yank

Re:Optical light?

[Anonymous Coward]

Um, I'm sorry you must mean British??? Most of the time when I hear torch I think fire.

Re:Optical light?

[dwarfsoft]

Dropping the superfluious 'u' because your accent doesn't allow you to actually distinguish between an '-our' and an '-or'. I think a lot has to do with accent, as most of the different spellings you are critical of actually serve a very good purpose, either in written English, or in enunciation.

Re:Optical light?

[dwarfsoft]

In which case, monitors were, and still are, large reptiles found on many islands in the Pacific (and probably elsewhere). In Australia 'cellphones' are called Mobiles also, pronounced mo-BILE. Yes, mo-BILE is also something potentially hung over an infants crib. This is yet another matter of context. If somebody here mentioned a cellphone the common answer would be "to whom? and for how much?". ;)

Re:Optical light?

[mysidia]

But a VFL is the right tool for the job, not a blazing unreliable Torch.

Re:Optical light?

[Anonymous Coward]

> There are a number of similar cases where homophones have different spellings in English but the same spelling in US English.

Homophones? Aren't they people who don't think words with the same spelling should be allowed to marry??

Re:Optical light?

[Anonymous Coward]

Your other three examples seem to be correct.

"telephone" is from 1835. Bell famously said "Mr. Watson, come here, I want to see you." in 1876.

Brits say "car" not "automobile", and "car" is from from around 1300

Re:Optical light?

[Hurricane78]

<with an accent> Hey, I’m check, you insensitive clod!

Re:Optical light?

[Anonymous Coward]

No doubt shortened from "electric torch".

Re:Optical light?

[Anonymous Coward]

And I'm pretty sure he isn't a Brit, he's made some mistakes in his arguments that indicates true ignorance of British English usage and I don't think they were deliberate ones.

Re:Optical light?

[tehcyder]

It works well for the Germans, after all, and they're much more famous for their engineering prowess and technology than the Brits are, so they're obviously doing something right.

Don't...mention...the...war...

Re:Optical light?

[tehcyder]

But if you're with a Brit and you're stuck in the wild at night with no light source handy, if you ask him to make a torch, he'll just be confused, and wonder where he's going to find a lightbulb and batteries out in the woods.

You win the now legendary "slashtard of the day award" hands down. Your prize is to spend a night stuck in the wild with someone from the UK with a pointy stick.

plug

[theheadlessrabbit]

Seems like this could be an effective plug for the analogue hole.

Cautious optimism should be shown. Sounds like something that could come back to haunt users.

Re:plug

[Kjella]

So how exactly is this different than DVI, HDMI and DP which are also digital? I think this has great potential...

Re:plug

[alvinrod]

It only makes it slightly harder to pirate something. If I can see or hear it, there's a way to record it. Killing off analogue is probably more about making you buy new versions of the same stuff more so than stopping piracy. I imagine some crafty bastard will develop a physical device that reopens the analogue hole for anyone who doesn't want to upgrade. The main issue is that the content companies will try to get such a device declared illegal, probably over piracy claims, but it's really about making everyone buy new hardware. Since the technical reasons for upgrading aren't overly compelling, it's necessary for them to do something to keep revenue up.

Re:plug

[Uksi]

Are you serious? How is this different from any other digital transfer method? And how could this possibly plug the analog hole? As long as you gots a speaker and a microphone or an analog transfer of audio (e.g. to your headphones), you have the whole. Tell me, is this going to do away with headphones?

Re:plug

[maxume]

You have to plug the analog hole on the record side. If you can see it, you can record it.

Re:plug

[forkazoo]

Seems like this could be an effective plug for the analogue hole.

Only if you misunderstand the analog hole. You can push it out as far as you like, but there will always be one. For audio, you eventually have to run analog voltages to a speaker element. If you can't tap that for some reason, you always wind up with actual *sound* that you can record with a microphone.

When the analog hole is pushed that far out, it can be damned inconvenient to make use of it, but it will always be there. Just having a protected stream of bits doesn't matter if you never present the media. Sure, it's secure, but nobody wants it if they can never play it. Once you try and present it, all your digital security stops and the media spills out into the real, analog world.

Re:plug

[Anonymous Coward]

The whole what? The whole WHAT!?? Oh please don't let me hang like this, finish your sentence!

Re:plug

[ceoyoyo]

Because we're all going to get a Light Peak socket in our heads?

Some concerns

[Anonymous Coward]

First, how much would it cost to produce a cable. And, second, how fragile would these cables be? Can you fold them up for storage or transport?

Are they practical?

Re:Some concerns

[frith01]

Fiber Optic cables have the standard issues of NO FOLDING/ CREASING, and minimal bend radius.

http://www.specialtyphotonics.com/knowledge_base/newsletter/0606/bend_insensitive.html

Re:Some concerns

[TheCycoONE]

Light Peak uses ClearCurve to circumvent the restrictions mentioned in the above article:

"In July 2007, Corning announced a new optical fiber known as ClearCurve that uses nanostructure reflectors to keep light trapped within the fiber even when bent around small-radius curves." - Wikipedia

(Notice your article is from 2006)

Re:Some concerns

[Anonymous Coward]

Last i checked, folding fibre is a fantastic pathway to buying a new cable.

Light Peak is going to fail horribly.
Your average computer tard folds cables in all sorts of stupid ways that destroy even copper wires...

Re:Some concerns

[Coren22]

Interestingly enough, fiber, even the glass stuff, can bend tighter raduis(i?) then solid copper wires. I have never actually had a fiber go bad from being coiled up, even when pulled straight and having a twist pull to a radius of my pinky finger.

Re:Some concerns

[izomiac]

They'll just en-sheath it in a metal coat that won't let you bend it beyond the recommendations. This coat will make the cable about 10 - 100 times more expensive. $49.99 seems to be what customers will pay, so they'll target that price point (it might be up to $75, I haven't looked at Best Buy's USB cables in the printer section lately). A lot of people still believe the "you get what you pay for" fallacy, so they'll think a $50 cable is spectacular. Speaking of which, Monster will come out with an EM-shielded, oxygen-free, gold plated version for about three times as much.

Re:Some concerns

[Anonymous Coward]

Judging by the looks of the picture in the original article, the cable looks coiled/wound up. So it would seem that Intel has overcome this issue somehow.

Re:Some concerns

[PitaBred]

Do TOSLINK (SPDIF) cables fail regularly? Are they prohibitively expensive?

Impact resistant?

[senorbum]

I wonder how resilient the cables are to, say, me stepping on one. Or even better, accidentally sliding the back of my couch into my desk and pancaking it.

Re:Impact resistant?

[saurongt]

Fiber is quite resilient to all sorts of stress except for tight bending. The cable can probably be reinforced to prevent that. I'm more curious why they use visible wavelength, when IR spectrum light experiences lower loss in fiber. Is it because the cable is not expected to be of great length, ala current USB cables?

Re:Impact resistant?

[DarkKnightRadick]

Wouldn't fiber be able to transfer both? I imagine they used visible light in the demonstration so people could see it transferring light. Last I checked, humans cannot see in IR or UV.

Re:Impact resistant?

[Lunix Nutcase]

Is it because the cable is not expected to be of great length, ala current USB cables?

Duh?

Vincent, we happy?

[myocardialinfarction]

http://www.youtube.com/watch?v=qH_tCxmug6E

Potential buyers?

[solevita]

The white USB plug in the demonstration laptop looks rather Apple-ish. Any rumours here?

Re:Potential buyers?

[Anonymous Coward]

Apple developed it that is why :)

http://www.engadget.com/2009/09/26/exclusive-apple-dictated-light-peak-creation-to-intel-could-be/

Re:Potential buyers?

[bjackson1]

The white USB plug in the demonstration laptop looks rather Apple-ish. Any rumours here?

More than rumors: http://www.engadget.com/2009/09/26/exclusive-apple-dictated-light-peak-creation-to-intel-could-be/

Re:Potential buyers?

[koiransuklaa]

Engadget story that neither Intel, Apple nor any other third party have confirmed. Isn't that pretty much the definition of rumor?

what is the killer app for it?

[alen]

i remember when Firewire got spanked by USB. Sure it was faster but no one cared if their ipod took an extra 10 minutes to sync the first time. not enough to pay the premium at least.

same here. printers are wifi these days. i haven't had a printer for years but will probably buy a wifi one soon just to print coupons from my iphone coupons.com app.
keyboards and mice can be had in bluetooth
since wifi is faster than the internet there is no reason to use this as a network cable
and how is it better than today's hdmi cables?

Re:what is the killer app for it?

[Locke2005]

Your printer can't print fast enough to keep up with WiFi, let alone Ethernet. The only application this cable makes sense for is the one they demoed on it -- HD video streaming. Think camera-to-computer or computer-to-project links. Now, if we just had a memory system fast enough to store and retrieve data at 10Gbits, we'd be all set!

Re:what is the killer app for it?

[DarkKnightRadick]

wifi also suffers from greater range limitation than wired. Also, wifi experiences greater packet loss then wired.

Re:what is the killer app for it?

[tonywong]

Not every printer is a consumer printer. Large format printers can get files gigabytes in size being thrown at them. There's lots of higher end applications that need lots of data, fast.

Re:what is the killer app for it?

[Tekfactory]

Backup, Data processing

I used to hang my tape backup off of my file server because it had the most data to backup and so the fastest interface to the tape drive was installed on the file server.

All of the other machines (and file server)were given Gigabit Ethernet cards, and attached to a Switch that could handle 2GB simultaneous per port. The file server and mail server then were bottlenecked by the speed of the hard drives and the tapes themselves.

We also had users on the high speed network that needed to process large segments of the company's data from the fileserver which usually involved reading it into memory processing it and writing back the changes. All of these little exercises would have benefitted from a faster bus speed on the motherboard.

We could have done some stuff with striping the RAID arrays and buying more memory for the SCSI controllers, today we could be caching most of the jobs to RAM on the desktops.

My biggest bottlenecks were the hard drives all round and users competing for the 1GB pipe to the Fileserver. Having some sort of 10Gbit interface on the File Server would bring it back to drives, but as cheap as Ram is today I clearly would have bought 32GB or RAM and cached the contested data on a Ramdrive.

Create a capacity and someone will find a way to use it.

Now all we did was market research data processing, I'm sure the 3D CGI movie folks could find a use for this on their renderfarms, and I wonder if there are uses in MMOs to increase the number of folks on a battleground or zone simultanouesly.

Re:what is the killer app for it?

[Locke2005]

If you've got a printer that can continuously print a gigabyte of data per second for extended periods of time, I'd love to see the paper handling system that it uses! How many full time employees are required to just feed paper into and carry paper out of the machine?

Actually, medical imaging is another application that does demand ridiculous bandwidth, but I wouldn't exactly call that "printing".

Re:what is the killer app for it?

[Real1tyCzech]

"wifi also suffers from greater range limitation than wired"

My wires must be bad, then. I can get a good 100ft from my router and still be connected, but if I go so much as a mm from my wire, I lose connection. That's around 100% packet loss!

Range from Wifi = 100ft.
Range from Wire = Zero.

Re:what is the killer app for it?

[Idiomatick]

I just pictured the bottom of the printer opening up and dropping dozens of ink cartridges onto the sheet of paper like a WWII bomber.

Re:what is the killer app for it?

[twidarkling]

I take it you've never actually seen a modern printing press, then, eh? You can actually get ones that take from a computer, rather than having to make plates.

Re:what is the killer app for it?

[alen]

i've been following the consumer computer market since the early 1990s. cheap and good enough always beats expensive and cool.

Re:what is the killer app for it?

[DarkKnightRadick]

Perhaps for you, but I don't like settling for "good enough" when I can get "best". Especially if best is cheaper than "good enough" (wifi is still more expensive than Ethernet).

Re:what is the killer app for it?

[DarkKnightRadick]

Range from Wifi router = 150ft indoors, 300ft outdoors
Range from wired router = 500 meters (1,640 ft, 10base5 coax) or 185 meters (607 ft, 10base2 coax).

Re:what is the killer app for it?

[Anonymous Coward]

Come on, people...

It was a joke. One can get farther away from a router than they can from the wire. ...get it?

Re:what is the killer app for it?

[DarkKnightRadick]

yeah yeah.

Re:what is the killer app for it?

[petermgreen]

wifi is still more expensive than Ethernet
Depends how you do your costing, getting a wired network proffessionally installed will almost certainly cost more than using wifi, especially if you don't want ugly cables on show.

Re:what is the killer app for it?

[DarkKnightRadick]

Perhaps, but the range is so much greater that, given the cost of many more repeaters vs. cable installation, I believe it would start to even out pretty quickly if you include installation which, rightly so, should be included in any estimate; my bad for not considering it earlier.

I still would rather have at least my critical systems wired instead of wireless. "Good enough" doesn't cut it when the box it's attached to is mission critical.

Patent and licensing situation?

[Anonymous Coward]

Does anyone know the licensing situation for this?

It would be awfully convenient if you could e.g. make a chip based on your own patents, which you license out, that provide a "standard" set of functionality, but when you pair it with your own hardware which is protected by patents you don't license out, it opens up an "experimental" set which soon morphs in regular language into a "full" set.

Horrible USB Connector

[Anonymous Coward]

Why did they have to stick it in the horribly designed USB connector?

The engineers responsible for that connector must have never made it past sophomore design class. You either make a part that is obviously asymmetric (d-sub, ieee1394, 8p8c) or one that is truly symmetric (RCA, TRS connectors). Instead, we're stuck with this symmetric-appearing but actually asymmetric USB connector that I try to plug in backwards half the time.

Re:Horrible USB Connector

[clone53421]

Instead, we're stuck with this symmetric-appearing but actually asymmetric USB connector that I try to plug in backwards half the time.

Who actually manages to plug it in correctly on the 2nd try? It usually takes me at least 3.

Re:Horrible USB Connector

[DarkKnightRadick]

I almost always get it on the first try, most of the time without looking.

Re:Horrible USB Connector

[sznupi]

I don't think it's that much worse from the "obviously asymmetric" - in case of those you have to usually look at the plug and the socket (and what if the latter is not clearly visible?) anyway. Plus USB connector typically should have convex logo on it meaning "up", and sockets should be in an orientation that makes sense for "up" (on laptops and hubs that's easy, even motherboards seem to comply...as long as you remember that "up" means "the side where all the cpu, memory and PCI slots are")

It has is strenghts too; grounding and mechanical properties being rather nice.

Re:Horrible USB Connector

[Anonymous Coward]

The right side up is always the side with the USB logo, or else the cable isn't compliant with the USB specification. You can tell even by feeling it.

Re:Horrible USB Connector

[Locke2005]

Plus USB connector typically should have convex logo on it meaning "up", and sockets should be in an orientation that makes sense for "up" My tower computers sit on the floor, with the USB ports 1" from the carpet. Since they are towers, the motherboards are vertical, not horizontal. So, at least in this case, I don't think "up" means what you think it means.

Re:Horrible USB Connector

[sznupi]

Could it be why it was in ""? (and why I mentioned how it ends up in practice for motherboards)

Re:Horrible USB Connector

[Idiomatick]

One of my PCs has the usb ports on the front of the case upside down. It confuses anyone that tries to use it.

Re:Horrible USB Connector

[twidarkling]

Unless the port itself is upside down.

Re:Horrible USB Connector

[Xoltri]

My personal average is 4 tries with 6 swear words. I've also learned that the USB connector fits convincingly in the network port, but interestingly enough it doesn't work in that port.

Re:Horrible USB Connector

[ClosedEyesSeeing]

It's the only rectangle plug that requires you to try it 3 different ways to get it to work.

Re:Horrible USB Connector

[Anonymous Coward]

Yeah, no kidding. Normally I'd peek at the cable connector. The "top" side has the USB graphic (looks like a plant or a mis-shapen trident). Making sure the USB logo was on top used to work, but recently I started seeing cables and PCs with the port inverted. I have a hub with some ports one way and the ones ones on the side are the other way. I can't tell why this is done, short of a design error.

Re:Horrible USB Connector

[Anonymous Coward]

Do not look into USB-shaped plug with remaining eye.
Even if you can't figure out its orientation.

Re:Horrible USB Connector

[Anonymous Coward]

I'd also like to mention that it first perfectly inside an 8p8c jack. I can't tell you how many times I've fixed a "broken" computer by plugging the USB cord in properly.

Re:Horrible USB Connector

[anethema]

Or sideways, like on every desktop

Re:Horrible USB Connector

[FreakyGreenLeaky]

We're not talking about a condom here, buddy.

Re:Horrible USB Connector

[RMH101]

I have often thought that if I totted up the amount of time I've spent attempting to plug USB cables in upside down, it would add up to several days of pure frustration. +1 to the asymmetrical design for a new connector..!

Re:Horrible USB Connector

[Anonymous Coward]

Why did they have to stick it in the horribly designed USB connector?

The engineers responsible for that connector must have never made it past sophomore design class. You either make a part that is obviously asymmetric (d-sub, ieee1394, 8p8c) or one that is truly symmetric (RCA, TRS connectors). Instead, we're stuck with this symmetric-appearing but actually asymmetric USB connector that I try to plug in backwards half the time.

More than half the time. It is such a pain. When will my netbook have mini-usb plugs on it?

Re:Horrible USB Connector

[DarkKnightRadick]

Funny, very funny! (:

Great

[Locke2005]

Great news for Lightfleet Corporation. They are now officially completely obsolete.

Re:Great

[DarkKnightRadick]

not really. Seems that Lightfleet Corporation could partner with Intel on the way to fully optical computers.

Master-Slave or Peer-to-Peer?

[lacoronus]

In regards to Light Peak replacing both USB and FireWire: Anyone knows if LP uses a hub-controlled topology like USB or P2P-ish like firewire? Even 100GB/s throughput won't do much good if we have a huge bottleneck in the hub. I've tried to find out but couldn't find anything in regards to this.

uh...

[grumpyman]

Integrating a HBA/HCA onto a laptop? What's new?

What do the British call real torches?

[maillemaker]

I'm curious: If the British call "flashlights" "torches", what do they call the a big stick with fire burning at one end?

Re:What do the British call real torches?

[Anonymous Coward]

Torches too. E.g. the Olympic Torch. If you think that's confusing, look up the number of different usages for the word "set".

Re:What do the British call real torches?

[Anonymous Coward]

Painful.

Re:What do the British call real torches?

[iammani]

The British refer to 'the a big stick with fire burning at one end' as a 'burning torch' or sometimes simply as torch..

You are welcome.

Re:What do the British call real torches?

[micilin]

And while we're at it: if the Americans call the "bonnet" of a car the "hood", what do they call a flap of fabric that goes over your head? (I kid, I kid)

Re:What do the British call real torches?

[Anonymous Coward]

Syphilis.

Re:What do the British call real torches?

[sznupi]

Well...are you certain they didn't use a "big stick with fire burning at one end"?

Re:What do the British call real torches?

[Jer]

I'd assume they'd call it a "torch" since the reason they use the word "torch" for what we call a "flashlight" is because it's a torch, just one powered by electricity instead of fire. Much like an electric oven in just an oven powered by electricity instead of fire.

A better question is "why did Americans decide it should be called a 'flashlight' instead of an 'electric torch'"?

Re:What do the British call real torches?

[Locke2005]

Because 1) It emits light, and 2) it flashes! Er, wait a minute... how about because it was the device of choice for use by nocturnal flashers, unlike torches, lanterns, and candles which had the unfortunate side effect of causing serious burns to precisely that area of the anatomy that the flasher most wanted to illuminate?

Re:What do the British call real torches?

[bluie-]

Ahahaha, mod up!

Re:What do the British call real torches?

[value_added]

From http://www.ideafinder.com/history/inventions/flashlight.htm

Late in the 19th century, many attempts to devise a portable electric lamp had been made, but the early ones were unsuccessful. Now a common household item, the lowly flashlight was once considered a novel toy. The first flashlight, or electric hand torch, was invented about 1896. Early portable electric lights were called "flash lights" since they would not give a long steady stream of light. The flashlights introduced in 1898 by Conrad Hubert's company, that would later become Eveready, were more trustworthy making Eveready the leading name in flashlights.

Note that in most other languages, it's called a varation of "lamp" or "lantern".

Re:What do the British call real torches?

[Anonymous Coward]

"A work that expires before its copyright never enters the public domain and thus enjoys eternal copyright protection."

That was sort of the point. It works well for corporations.

Re:What do the British call real torches?

[Anonymous Coward]

Because early flashlight had horrible battery life, and would have less than a half hour of runtime. Therefore, one would need to flash them on and off periodically to see where they were going, else they would be dead before they got there.

WOOHOO! I can watch tele vision

[Colin Smith]

That is seeing things transmitted from far away!

Amazing!

One cable to rule them all...

[rthille]

For a desktop, I don't care about this. For my laptop. Especially for my Mac laptop, I want this because it's the closest Steve will get to doing a 'dock' for a Mac Laptop again. (I owned a Powerbook Duo 210/2100, so don't feel the need to remind me about them).

For now when I put my Mac laptop on the stand on my desk, I plug in 6 cables: power, network, display, Firewire, and two USB (one is for the KVM, so that can't go into a hub, and the KVM is only 12Mbit, so the other cable can't go into it for reasonable speed).

With Light Peak, I could do two cables: Power and Light Peak.

Re:One cable to rule them all...

[harrkev]

You do realize that you can reduce this. You can buy a "docking station" that has a USB hub, USB ethernet, and USB audio on one box. Then the KVM can go into that for the mouse/keyboard. You will still need power, firewire (who uses that anymore) and video. But still, that would eliminate two cables.

There are also USB docking stations that include their own graphics chips, but those have their own problems.

Re:One cable to rule them all...

[Anonymous Coward]

USB ethernet isn't real ethernet....

Power? FireWire ,enet, USB give power does this?

[Joe The Dragon]

Power? FireWire and USB give power also E-net can give power does this?

Re:Torch into one end with two little dots of ligh

[iluvcapra]

I'd be curious to see how many people are old enough to actually get this one.

Short sightedness

[dreamchaser]

It's nice they've developed a way to data on plastic discs, but it does the average user no good as long as we're using magnetic drums and punch cards. I guess these so called 'floppy' disks are aimed at the high end workstation market, then?

(end sarcasm)

Re:Short sightedness

[Anonymous Coward]

Or in other words: cease development of everything that is anything other than an infinitesimal improvement over a mature technology.

Or in other words: cease basically all development that requires people with more than 3 years of college.

Yawn...

[Anonymous Coward]

Yawn...

I used to be the sysadmin for a high school that had a multimode fiber-to-the-desk LAN installed in 1994. It's still running at 10 mbps, but that exact same wiring carried 1 gbps to outbuildings, and could carry 10 gbps if so desired.

Re:Power? FireWire ,enet, USB give power does this

[Anonymous Coward]

The data channel is optical but they are planning to include copper lines and pins for power however no specifications for how much voltage/amperage have been release (at least to me knowledge).

Re:Torch into one end with two little dots of ligh

[dskzero]

It isn't really all that old, to be fair.

Re:Torch into one end with two little dots of ligh

[Amouth]

we are on /. i'd hope most people here got it old or young..

Does not replace, it bundles!

[chaim79]

Something to remember as you look at this, the LightPeak connection isn't just a connector onto itself, it's also designed to handle all other connector types (eSATA, USB, Firewire, DVI, etc). It's designed to be the one port you plug into your laptop while at the other end a dozen different devices are connected to it, all using different protocols.

Contradictions...

[courteaudotbiz]

There's almost no limit to the bandwidth - fibres can carry trillions of bits per second

In the same sentence, they say there's no limit, and that the limit is trillions of bits per second. This means, using short scale number naming, 116 Gigabytes per second for 1 trillion bits per second.

Pretty high, but contradictory in itself.

Re:Contradictions...

[comm2k]

You quoted it yourself yet somehow managed to miss the almost part?

Re:Contradictions...

[thanasakis]

1 trillion bits /sec, 1Tbps? Not so fast. At this point, optic fibers can carry multiple wavelengths, each carrying up to 100Gbits/sec. Mind you, 100Gbits are pretty recent, with implementations coming up from the major companies, like alcatel-lucent (plz correct me if there's already a commercial product). A more sane number is 40Gbits, 10Gbits or even 2.5Gbits. To multiplex and demultiplex multiple lambdas in a fiber you need expensive optical de/multiplexers, an amount of pairs of transponders equal to the number of your lambdas and possibly amplifiers/regenerators. Although it should be easier if we are talking about fibers a few meters long, the electronics and the optical components to handle these rates are neither easy nor cheap.

It would sound more reasonable to expect to have a single connection maxing out at 10Gbps, preferably a 10Gbit ethernet connection.

Re:Power? FireWire ,enet, USB give power does this

[TheCycoONE]

The plan is to include a copper wire along with the optical wire for powering devices.

It is sorta mentioned here: http://news.zdnet.com/2100-9595_22-346181.html

Re:Torch into one end with two little dots of ligh

[iluvcapra]

16 years ago. Back then Apple was selling Quadras for $5000 and office managers were buying Windows 3.1 for Workgroups. The FDIV bug was discovered months before the Linux kernel 1.0 was released, and people still regularly used something called "Grolier's Encyclopedia" on CD-ROM to watch 320x240 15fps movies of the Apollo launch. Phil Hartman (God rest his soul) was selling Phillips CD-i players, A kid in my neighborhood had just bought a JVC X'EYE, and Conan was still writing for Simpsons.

Flash light -- name origins

[becker]

The word "flashlight" was actually a derogatory term. The carbon-zinc batteries developed insulating bubbles under high load, and the light flickered and flashed. I'm sure most readers here have experienced this -- imagine how bad it was when the technology was new and barely understood.

That said, I have to agree: torches are what the mob from the village carries when they come to break down the door of your lab.

Re:Flash light -- name origins

[Grishnakh]

The word "flashlight" was actually a derogatory term. The carbon-zinc batteries developed insulating bubbles under high load, and the light flickered and flashed. I'm sure most readers here have experienced this -- imagine how bad it was when the technology was new and barely understood.

According to Wikipedia, it was called "flashlight" because the carbon-zinc batteries couldn't deliver current for long, and had to "rest" periodically:

These early flashlights ran on zinc-carbon batteries, which could not provide a steady electric current and required periodic 'rest' to continue functioning.[5] Because these early flashlights also used energy-inefficient carbon-filament bulbs, "resting" occurred at short intervals. Consequently, they could be used only in brief flashes, hence the popular name flashlight.

Of course, it sounds like the Wiki article is putting a more positive spin on it; your description sounds like what really happened when you left them on too long.

That said, I have to agree: torches are what the mob from the village carries when they come to break down the door of your lab.

Yeah, that's like what I was thinking too: mobs of villagers looking for werewolves or somesuch, carring torches and pitchforks.

Re:Flash light -- name origins

[petermgreen]

Yeah, that's like what I was thinking too: mobs of villagers looking for werewolves or somesuch, carring torches and pitchforks.
Torches like this? ;) http://www.wickedlasers.com/lasers/The_Torch-74-41.html

Re:Flash light -- name origins

[wagnerrp]

According to Wikipedia, it was called "flashlight" because the carbon-zinc batteries couldn't deliver current for long, and had to "rest" periodically:

Wait, is this why flashlights in nearly every video game behave this way?

I was there when they started working on this....

[pdxp]

And their only reason for making it is so THEY can specify the required DRM compliance for high-bandwidth devices. It was the foremost issue during development. It's the only reason they've re-branded and protocol-ized existing technology. HDCP 3.0 anyone?

12mm square chip

[Anonymous Coward]

Does it mean the chip is 12mm x 12mm square or that it's actually 12mm^2 area? There's quite a huge difference there..
I can't imagine how you could dedicate that much space to the lightpeak chip, unless it can handle multiple streams.

Re:Torch into one end with two little dots of ligh

[dskzero]

I just tend to remember that a lot better by reading the Quake readme.txt, and knowing that the game used floating point stuff. I had no idea what that was. (keep in mind my natural language isn't english)

What's wrong with 10G optical Ethernet?

[Ungrounded Lightning]

Why a new proprietary optical transport when there are already standards-based transports that do the job just fine?

Or is this just a cheap, short-range, optical ethernet transceiver with a new connector, cabling system, and optics-integrated interface chip?

Two fibers would be consistent with using integrated LEDs for transmitters rather than separate lasers and/or using two frequencies to go bi-directional on one fiber. For short range you don't need coherent light or single-mode fiber.

Re:What's wrong with 10G optical Ethernet?

[bill_mcgonigle]

All of the above. Cheap, integrated power and they also use a coating on the sheathing so that when you bend the cable to a small radius your monitor doesn't go blank. Imagine being a tech support rep if they use standard single-mode for this application.

Or "one cable to rule them all" if you'd rather.

Re:What's wrong with 10G optical Ethernet?

[Hurricane78]

No lock-in.

Remember how Firewire fell into oblivion on the standard consumer desktop, because of USB, despite being technically far superior?

USB = power to intel.

two separate HD video streams? ahahhahah, wow

[dnaumov]

The very highest quality 1080p bluray rips have a bitrate of about 18Mbit. Meaning as long as you can handle 2MB/s sequential read speed, you are fine. Double this to 4MB/s for 2 streams. How is 4MB/s supposed to be impressive? Even the worst, cheapest USB flash sticks are capable of over 10MB/s. If Light Peak is supposedly capable of 10Gbit both ways, why on earth is it demoed with this?!

Re: two separate HD video streams? ahahhahah, wow

[m85476585]

That's compressed size. Uncompressed HD video is gigabits per second, and most displays take in uncompressed video.

Re: two separate HD video streams? ahahhahah, wow

[Anonymous Coward]

An HD video stream (1080p60), as seen by HDMI or DVI is 1920 * 1080 * 24 * 60 = 2,985,984,000 bits per second. That's almost 3Gbps without even taking into account framing or time between frames. USB 3.0 can nominally achieve that bit rate, but it can't come anywhere close to that bandwidth. Of course if they were running 2 HD displays, they could be using up to 6Gbps. Even SATA can't achieve that kind of sustained speed.

dom

Re: two separate HD video streams? ahahhahah, wow

[Anonymous Coward]

The very highest quality 1080p bluray rips have a bitrate of about 18Mbit.

Actually, they average better than 30mbps. My mistake... you were talking about the re-compressed garbage you pirated off some torrent weren't you...

So this operates on a demo machine...

[sarkeizen]

...but I can buy an actual 10Gb copper adapter for a machine I have now. Granted that they've come up with some spiffy ways of using fiber in a consumer cable (as opposed to safely ensconcing it in conduit) and a 10Gb copper NIC is prohibitively expensive but that said it's not unreasonable to believe that the price of copper 10Gb NICs will drop before this thing is available in consumer products and although I get that this is essentially a point-to-point connection to replace things like USB. Yet Ethernet still seems - at this point anyway - as viable a choice as some new thing...which we will have to use in addition to Ethernet.

Re:So this operates on a demo machine...

[Bruha]

You will not see consumer 10G Nic's, they're worthless in datacenters, thus no real way to push costs down. You can only get 10G copper about 100 feet with any reliability and it's spotty at that. Best to stick with fiber lines, best investment with 50 micron cables that will last well into 100G.

Re:So this operates on a demo machine...

[drsmithy]

You will not see consumer 10G Nic's, they're worthless in datacenters, thus no real way to push costs down.

Er, what ? Who _wouldn't_ be using 10G in new hardware ? *Especially* given the growing prevalence of virtualisation, iSCSI and the emergence of FCoE ?

You can only get 10G copper about 100 feet with any reliability and it's spotty at that.

So, basically perfect for the typical case of 5-10 racks full of servers connected to each other and then uplinked to the core ?

Re:So this operates on a demo machine...

[Anonymous Coward]

I'm not sure what you're thinking of but I've seen deployments triple the distances you are talking about on Cat7. As 10GBASE-T prices drop I'm willing to bet you *do* see them in the datacenter. People use 10GSFP+ which has a tenth of the distance.

Re:So this operates on a demo machine...

[sarkeizen]

...and I just checked GigE premiered at $800 per port. Almost double the price of a current 10GBASE-T NIC. Seems like it has a shot.

Re:Torch into one end with two little dots of ligh

[TheRaven64]

And a 60MHz Pentium was fast. I remember wondering what people would possibly do that required a Pentium, when a 66MHz 486 was already really fast. Then Windows 95 came out.

Re:Torch into one end with two little dots of ligh

[Anonymous Coward]

Yeah, we LOL'ed about it then. This was back when the PowerPC was looking to be real competition and Motorola was still in the processor biz.

Can this tie in to ATI and NVIDIA cards?

[Joe The Dragon]

Can this tie in to ATI and NVIDIA cards?

If locked to Intel video this will die fast.

High bandwidth is good..

[Madsy]

But low latency is just as important. Why is it always the bandwidth which is advertised in big numbers and not the latency?

Re:High bandwidth is good..

[Anonymous Coward]

But low latency is just as important. Why is it always the bandwidth which is advertised in big numbers and not the latency?

Because, on a point to point link that is streaming anywhere near 10Gbps, the latency is pretty much guaranteed to be negligible in the grand scheme of things.

Even if it uses huge 64kByte frames, and the switching logic buffers 2 frames, the latency will only be around 100 microseconds.

The really useful information isn't just a raw latency figure, it's the efficiency of the link in maintaining the maximum data rate without the receiver having to assert flow control while it processes the incoming data.

For any stream, the latency is set by the data rate, the frame size, and the number of frames in the pipeline.

so this is 10GBASE-SR

[WhiteDragon]

It seems we already have a standard for 10 GBit/s over short lengths of fiber: 10GBASE-SR

Intel has reinvented fiber optics

[Anonymous Coward]

Sounds like Intel has reinvented fiber optics. So what's the big news?

AntiInformative

[Gary W. Longsine]

Congratulations, you have been granted the Bogon of the Day Award. You have stated plainly counter-factual claims, which could have been easily verified if you had spent only a mere few seconds of your life checking in with the Google, prior to posting. Furthermore, the fact that you were subsequently up-modded to "Informative" means that you stated your claims in a sufficiently compelling manner as to fool the hard working Slashdot moderators.

Others have or will point out the errors in your claims, or you could consult the Google. I merely present this award, and once again offer congratulations. It's a special thing, to construct the most idiotic post on Slashdot, in a given day.