USB On-the-Go Go Go Go

abhikhurana writes " There is an interesting article on CNET about a new USB standard called USB On-the-Go. Apparently this new technology is an offshoot of USB2 and it can remove the limitation of the master slave operation of normal USB devices, where you need a Host PC (the Master) to talk with the peripherals (the slaves). So using this, theoretically you can print using your digital camera directly on your printer or maybe connect two PDAs together to exchange some files. One thing that the article doesn't mention though is the speed one can expect from such a connection. If its as fast as USB2 then I think it can also act as the replacement for NICs for interconnecting two PCs. But considering that many wireless technologies like bluetooth offer similar opertational capabilities,albeit they are much slower, can USB On-the-Go really be a success? "

Comments

Not a chance..

[Loki_1929]

Nothing can compete with the power of this new and exciting technology that's about to take off, called... infrared !

FireWire already Goes Goes Goes

[Anonymous Coward]

FireWire works in a peer-to-peer fashion. You can hook up components without a computer to mediate.

Re:FireWire already Goes Goes Goes

[HakuMage]

yes, very good feature, but why nobody use it ? :(

it's like firewire network. if i remember well, there is only one company which made available necessary softwares ...

It will be great if for example, home users can made all their computer or domotic stuff with only one or two interfaces. (not really the case in enterprise :)

Re:FireWire already Goes Goes Goes

[terminal.dk]

Yes, it is yet another playing catchup on Firewire. I winder what the big fuzz is all about. Old technology re-implemented, and sold as new.

Windows XP even includes TCP/IP drivers for Firewire, and will allow you to network machines using FireWire. It finds both the FW interface on my SB Audigy Soundcard as well as my stand-alone FW card.

People should skip that USB crap, and go to the source, where it is proven technology with years behind it.

Re:FireWire already Goes Goes Goes

[Zathrus]

People should skip that USB crap, and go to the source, where it is proven technology with years behind it

Sure, except that my printer, my scanner, and my camera already speak USB. I suppose I should just throw them out and buy new stuff.

Hrm... where are those cheap firewire ink jets? Or scanners? And, ya know, I just bought this Nikon 770 a year ago. I think it's still good.

People buy the interface that works with their components. Frankly, every PC shipped in the past 4 years has at least USB 1.0 on it. Relatively few have Firewire/IEEE1394. And since Firewire is more expensive to implement than USB, you can count on the vast majority of devices to continue implementing USB and ignoring Firewire.

Does Firewire have its place? Sure. But it's not on most consumer devices. Up until USB 2.0 it was the only choice for devices that needed high speed digital data ports (like video cameras), but USB 2.0 is still cheaper to implement.

Firewire isn't going to die off by any means - it's solidly entrenched in the video market, and HDTV is likely to make this even more true. But lay off the "USB sucks, Firewire r0x0rs" - USB does very well for a very broad selection of products and at a fraction of the cost.

As for the people whining about USB sucking CPU cycles - uh... and you're telling me that you max out a 1 GHz+ CPU constantly? Gimme a break.

Re:FireWire already Goes Goes Goes

[Anonymous Coward]

It becomes a BenzeneWire device.

Re:FireWire already Goes Goes Goes

[bogado]

I don't think the people that he talked about was the consumers, but the hardware vendors, why there is no motherboard with firewire? Why all the consumer eletronics are USB, since firewire is much better?

Re:FireWire already Goes Goes Goes

[JebusIsLord]

Nah, in Windows XP at least the firewire port on my soundblaster audigy automatically showed up in network connections with tcp/ip enabled. nothing proprietary. And lots of devices use that feature, like digital video cameras that can wire pictures directly to a firewire-enabled printer without the use of an intermediary PC.

Re:FireWire already Goes Goes Goes

[Cinematique]

Yeah, but FireWire isn't USB. *cough*

Personally, I think someone should come out with a standard called WhiteJaw. I envision it as having the capability of wirelessly communicating with various accessories such as cell phones and PDAs. I see a potentially HUGE market for it.

Someone needs to get on the ball and make something that'll help computers and printers communicate back and forth, too.

Re:FireWire already Goes Goes Goes

[Otterley]

The answer is easy: USB functionality is built into the south bridge of many chipsets, starting with Intel's back in 1997. Intel is the primary backer of USB.

Firewire is fairly easy to add -- however, because few if any south bridges have it, it requires a separate chip to be added to the motherboard, and the chip adds $5-$10 to the cost of production. This corresponds to a $20-$25 increase to the retail price -- and if your competitors aren't doing it, it's understandably difficult to justify doing it yourself.

Re:FireWire already Goes Goes Goes

[gsfprez]

Peripheral Deivce Maker A: "But FireWire is a proprietary standard....and as we all know, Apple is all about proprietary standards and high-priced hardware.. not about interoperability.

And now, finally, here in 1996, Apple computer is about to reap what they sew because they are on the verge of goin... (shut up.. i'm talking).... going out of business... Beleagured App... (what do you want?)

Oh... i was in a coma? What year is it? 2002!? Jobs came back!!!? Macs now use the same memory, hard drives, PCI slots, VGA monitors and USB devices that are in PC's?

And most PCs now have USB and Firewire built in? Oh.

shit."

i think that many peripheral device manufacturers still don't get the fscking clue... that's why they went along with Intel and USB 2.0 - and now, there's a bunch of cables and hubs and devices that just confound "Your Mom" when she wants to buy something in the store... the plugs look the same, but now, this will be a 3rd standard on the same plug - so it will be even MORE confusing to "your mom" as to what cable can go where.

USB 2.0 has no purpose - other than to save face for Intel and to give Apple a cold shoulder for their perfectly good FireWire.

Re:FireWire already Goes Goes Goes

[Zathrus]

You mean like the Asus A78VX, Abit AT7-MAX2, Aopen AX4BMAX, Gigabyte GA-8IEXP, Soyo SY-P4I Fire Dragon, or Asus P4B533-E?

These are all new motherboards, and most of them are on the high end of pricing.

Firewire is more expensive to implement, period. And the number of devices that can substantially benefit from the faster speed of Firewire are very few (basically digital video... high end digital audio as well, but that's so stratospheric as to be irrelevant in the consumer market).

Re:FireWire already Goes Goes Goes

[Zathrus]

You forgot UD: Cancer Research which is what I run.

And, frankly, if you have so many spare cycles to run a distributed agent, you have the cycles to spare for USB. It's below the noise level.

Re:FireWire already Goes Goes Goes

[TheToon]

>I think someone should come out with a standard called WhiteJaw

No no... not WhiteJaw. It's a bit racist to imply that only white people can use it. And Jaws was a terrifying movie that will scare the kids.

BlueTooth is a much more PC name...

Re:FireWire already Goes Goes Goes

[utexaspunk]

Sure, except that my printer, my scanner, and my camera already speak USB. I suppose I should just throw them out and buy new stuff.

*News Flash* - your printer, scanner, etc. may speak USB, but they don't speak USB 2.0, or USB-on-the-go. So they're not going to be able to take advantage of this. Why invest in an interface that has to keep coming out with new standards just to keep up with one that already supports all the functionality you want?

Besides, consumer devices ARE moving to firewire. Look at the iPod, or a lot of new digital cameras. Everyone's eventually going to have FW because eventually everyone's gonna want to play with DV. Apple was smart enough to recognize this. Apparently you aren't.

Re:FireWire already Goes Goes Goes

[gpinzone]

Does each firewire device get assigned a unique TCP/IP address? What if you don't have DHCP installed?

Re:FireWire already Goes Goes Goes

[gpinzone]

They probably would have if someone would have simply made a Firewire - USB hub. How else are you gonna plug that Intellimouse Explorer or USB keyboard into the firewire port? Sad. Yet another superior technology getting burned because of a lake of proper marketing.

Re:FireWire already Goes Goes Goes

[Mike Hicks]

Well, I don't know how many of these "USB On-the-go"-capable devices you'll need to get in order to use it. I suppose you'll only need one thing (could be a hub, maybe even a "smart" cable), but you need something there to be a master device.

Re:FireWire already Goes Goes Goes

[ncc74656]

I don't think the people that he talked about was the consumers, but the hardware vendors, why there is no motherboard with firewire?

We've got an Asus P4S8X on order at work that includes onboard FireWire (and USB 2.0 and Serial ATA, as well). Other SiS 648 motherboards ought to have it as well, since FireWire is built into the chipset.

Re:FireWire already Goes Goes Goes

[Rader]

I would be happy with the cure to cancer no matter what the cost. A cure is a cure, and the information will dissiminate.

Besides, chemo can't be very cheap. And dying isn't very profitable.

However, your points are very valid. It would be NICE if they spent some front-time on their goals and methods like SETI so clearly did back in the beginning.

Re:FireWire already Goes Goes Goes

[bogado]

The only problem for them is that people will someday realise that AMD is faster and cheaper. :-)

Re:FireWire already Goes Goes Goes

[Ryan Amos]

Yeah, but firewire wasn't designed to run things like mice and keyboards, which USB was. I also doubt there are many firewire printers out there (at least in the low price inkjet category.) Don't get me wrong, firewire is great, but it's not the end-all solution to everything. Also, Apple has firewire because they helped develop it :) Apple gets royalties from every firewire device sold, and as such, it's in their best interest to push it. Granted, FireWire is the superior technical standard for high bandwidth situations, but USB is cheaper and more than adequate for say, a printer, a Palm, an input device, etc.

what about IEEE1394x

[radoni]

isn't this a perversion of what USB was designed for... keyboards and mice? ;)

USB for mice, IEEE1394x for men.

Developer Info

[MountainLogic]

More On The Go details can be found at the USB association's web site

Re:Developer Info

[__aasmho4525]

another location i found a good bit of information:

http://www.semiconductors.philips.com/buses/usb/pr oducts/otg/index.html

How many more USB do we need?

[euxneks]

next up:
USB3
USBSEEME
RUSB (Are USB?)
USBT (U Suck Big Time)
USBX! (for X-box!)
USBPS2 (for the mouse!)

I thought that USB was so we wouldn't have this many connections??

pics from my camera straight to my printer?

[garcia]

what if your camera makes insanely large photos and you would prefer to have them at 1024x768 or smaller?

I print out, maybe, 10 pictures a year. The rest are for going on the web.

I have no real desire to immediately print out a picture. But I suppose if you were going to have a stereo component that was MP3 capable, plugging your iPod (or whatever) directly in and having it transfer would be nice.

Re:pics from my camera straight to my printer?

[ncc74656]

what if your camera makes insanely large photos and you would prefer to have them at 1024x768 or smaller?

just change your dpi. 1800x1200 at 300 dpi is a standard size photo. odds are your camera won't have enough MP to print a full sized 8.5x11 sheet of paper (without interpolation).

Unless you have a dye-sublimation printer, you're not printing your photos at 300 dpi anyway...odds are you're closer to 100-150 dpi. Since most printers can only squirt ink or not squirt ink (there's no software-based control of how muck ink is squirted), you need to use dithering to get 4 (or maybe 6) colors to look like 24-bit color.

can USB On-the-Go really be a success?

[stratjakt]

Absolutely.

There are lots of advantages over bluetooth, etc.

No batteries, you can power stuff off the USB inteface.

Wireless (in)security.

Interference.

Cheaper.

Sometimes wireless stuff is just a pain in the ass.

It'd be nice to be able to just buy a digital camera and a photo printer, and be able to bypass a computer altogether. Not every electronic device in your home need be linked together somehow.

The 'interface' aspect of just plugging something in to 'connect' it is easier for the layman to grasp than having devices announcing themselves to each other over the air, etc.

Re:can USB On-the-Go really be a success?

[halftrack]

Agreed(ish), but ...

No batteries, you can power stuff off the USB inteface.

When it's a direct connection, does the electricity magically materialize in the wire.

(Sorry for the trollish formulation.)

Re:can USB On-the-Go really be a success?

[Locutus]

Remember that there are companies which HATE Bluetooth with a passion. A light-weight wireless protocol is a boon to device manufacturers but one particularly large software company finds that enabling the world is a very bad thing.

Jumping on a new wired standard now and with a savings of POSSIBLY only a couple of bucks is silly. This could be Intel realizing that USB is going away and Bluetooth taking over it's role( with FireWire for the high end ) but that's life. That's progress. I would still look to see if Microosft isn't mixed up in this somehow since this seems to be a shot at Bluetooth and they don't want to see Bluetooth do things like enable Palms and or enable other non-bloated non-WinCE based handhelds. Enabling these non-WinCE devices pushs Windows aside and that's not what innovation at Microsoft is about. IMHO.

LoB

Re:can USB On-the-Go really be a success?

[foobar104]

Sometimes wireless stuff is just a pain in the ass.

Only if you're not using the right operating system. With OS X, wireless ethernet and Bluetooth both work pretty darned near perfectly. Wireless ethernet requires no set-up at all, and Bluetooth only requires "pairing" your devices (or whatever it's called; I don't have my Bluetooth gear in front of me right now).

Wireless stuff is not a pain in the ass. Some wireless implementations are a pain in the ass. Don't confuse the two.

Re:can USB On-the-Go really be a success?

[stratjakt]

OSX? this is about device to device connections, without the host master (your computer).

I don't see any wireless printers that automatically know when a wireless camera is close, and start printing. It'd be easier and cheaper to just plug 'em in.

Re:can USB On-the-Go really be a success?

[stratjakt]

> When it's a direct connection, does the electricity magically materialize in the wire

http://www.allegromicro.com/techpub2/usb/power.h tm

So the printer maybe need be plugged in, the camera/device need not be. Depends if the portable need more than 100mA or not. (Maybe a camera would to take photos, but not to dump its contents out).

Re:can USB On-the-Go really be a success?

[foobar104]

Well, let's see. I can print to my Bluetooth HP printer from my Bluetooth Palm. They just see each other and start printing, basically. (For obvious reasons of security, you have to explicitly pair the two devices as part of a one-time setup.)

What else? My laptop automatically sees my Bluetooth T68, and makes it available to me for lots of things, not the least of which is dialing out to the Internet over the T68 like you would a modem.

And, while I don't own this gear myself, my friend from Australia has a Sony video camera with Bluetooth, and he says he can print stills from it on his Bluetooth-equipped printer. Again, I don't own that stuff, so I won't swear to it. But he says it's neat.

It'd be easier and cheaper to just plug 'em in.

Cheaper, definitely. Bluetooth-equipped stuff is expensive, at least right now. But easier? No way. I haven't carried a cable with me, except the FireWire cable for my iPod and the power adapter for my laptop, in months. No more USB cables for printers or my Palm or my digital camera or any of that crap. It all just works.

Strange

[Jobe_br]

I was under the impression that Intel had purposefully designed the USB protocol to be processor bound. This type of connectivity is already provided by Firewire, so I don't particularly see why this would be beneficial, unless devices somehow don't need to be explicitly USB On-the-go compatible or (more likely), the chipset/firmware for USB On-the-go is cheaper to produce/license than Firewire is.

In any case, my chips are still on Firewire - its a solid, fast and proven interconnect technology. With transfer speeds in excess of 38MB per sec. (76% of theoretical max of 50MB/sec) - I'd say they're doing quite decent. I'm not sure what USB2 is up to these days, but last I heard, they were still a far cry from their goal of even being faster than Firewire, in real world applications.

Incidentally - I don't mean to start a flame war on the benefits of Firewire v. USB - so don't get started. The transfer speed I threw out above is a valid benchmark for a external RAID array (that has drives fast enough to support that transfer rate and a equivalent RAID configuration to boot). I don't follow USB2 developments closely, so if I'm mistaken on its real-world speeds, forgive me and don't waste /. bandwidth by flaming me :).

Cheers.

Re:Strange

[Yohahn]

The USB PCI controler cards can directly access memory, like many other PCI cards
The PCI interface in practice (not theory) is a bottle neck here. USB2/Firewire on the motherboard chipset is always going to be faster than on the PCI bus.

Where did the USB is processor bound rumor come from?

If I'm wrong, will somebody please supply a source so I can correct myself?

Re:Strange

[Yohahn]

Here's an article to back up my claims that the speeds are similar. This is not a conclusive comparison, but it will show that USB 2.0 is comparible in speed to Firewire IN PRACTICE.

Re:Strange

[Yohahn]

I'm pretty sure you are wrong.

I'm pretty sure interrupts are part of USB.

Try doing a grep for "interrupt" in the "drivers/usb" directory of the linux source.

Can you cite your claim?

Re:Strange

[Yohahn]

Newer chipsets are building it into the chipset, not having to go through a bridge. You're probably looking at a motherboard using the NCR host adaptor

As put by the poster "DeadSeaTrolls" here.

"The current generation of USB2 PCI implementation can't saturate the bus, but those built on to the chipset start to get a lot faster (ICH4/nForce2)."

Re:Strange

[Yohahn]

As I said it wasn't conclusive.
But the ultimate authority in the end is what really happens. Copying large files really happens, and sometimes, it's good to just time it.

That said, I'd like to see a more worthwhile test, anybody seen one?

Firewire : Same Price, Twice the Speed

[shking]

By the time this ships, Firewire will be twice as fast. Firewire has always had this capability & is about to double its speed. To quote the article:

The debut of USB On-the-Go helps USB 2.0 draw closer in functionality to rival IEEE 1394, also known as FireWire. The 1394 specification, which is used widely in consumer-electronics devices such as digital videos, also has so-called host capabilities. Analysts say it would cost about the same to implement 1394 and USB On-the-Go in devices.

Re:Firewire : Same Price, Twice the Speed

[cdrudge]

You can go to the same store and buy a Firewire card for about the same price. And for that price, you can get the ability to sustain 400 Mbit transfers where USB 2.0 can only burst to those speeds. FireWire has also been around longer and I think is more mature then USB 2.0. Firewire 2.0 will come eventually, but right now, I think that it still beats USB 2.0 as a technology

Re:Firewire : Same Price, Twice the Speed

[windex]

My TiBook, my new Gateway Laptop, my Desktop, and my digital camera equipment.

Anything else?

FireWire am > USB 2.0.

Re:Firewire : Same Price, Twice the Speed

[Elwood P Dowd]

Firewire has been about to double its speed since it was invented.

The more relevant point is that 400mbit firewire already transfers much faster than 480mbit USB 2.0, simply because it stays closer to its capacity longer.

The reason that USB On-the-Go will be more desireable than firewire is that manufacturers can be sure that their customers will be able to use the USB On-the-Go devices. With firewire they will not. For the near term at least, I'm sure this will trump any speed concerns.

Re:Firewire : Same Price, Twice the Speed

[windex]

I did NOT SAY Firewire 2.0, I said *Firewire*. No USB 2.0 implimentation I have ever seen can match Firewire's throughput. End of story. People use whats better, not what claims to be better. If they didn't, everyone would run Windows for everything. :)

Re:Firewire : Same Price, Twice the Speed

[swb]

I think this is an excellent argument for shipping something functional but not perfect in volume and fixing it over time versus spending a long time developing a more perfect version of it.

How hard would it have been for Intel to have done USB2 the first time? Probably not that hard. How long? A lot longer. USB support would probably be a "new" thing in Windows *XP* or Win2k SP2 add-on if we had to wait for them to make what USB2 delivers in a USB1 release.

By the time USB3 is out, it will probably be exceed Firewire 2 in every respect because they've had time to iterate over it instead of trying to make it perfect the first time.

Don't get me wrong, I hate to see non-functional systems released but sometimes doing *something* part-way is better than doing nothing while you try to make it perfect.

Re:Firewire : Same Price, Twice the Speed

[cdrudge]

The specs for USB 2.0 specify 480 Mbits/sec. Firewire specs are 400 MBits. The difference is that the USB can not substain the high transfer rate. I believe the reason was that USB is a slow protocol. It wasn't designed originaly to transfer large amounts of data quickly.

If a USB 1.x device is present, then the transfer rate is further reduced. USB also has the whole master/slave thing. The master device (usually the computer) has to control the flow of data. This is what the USB On-The-Go will help with.

The PCI bus wouldn't be the bottleneck anymore then what an on-board chipset would do. Most likely the on-board would just be hardwired into the PCI bus. , much like sound/video cards are these days. The slowest PCI busses have about 80-90 MBytes of available bandwidth after accounting for system overhead.

There usually isn't anything that is truely magical about having something built in on the motherboard. The main reason for doing this is cost. The chipset manufacturer can incorperate all the different interfaces (USB, IR, IDE, etc) into one chip cheaper then what it would cost to produce individual chips for all the above, the board space, connectors, etc.

This is one reason why USB has taken off a lot quicker then Firewire. Intel supported USB initially, so it included it in it's chipsets. Firewire wasn't "theirs" so they don't include it. Guess what chipset is used for many motherboards.

Re:Fireware already does this...

[SlamMan]

I think there's a firewire-usb bridge you can add on to use up a firewire channel and take some of the load off your usb ports.

But I hear you. It took iomega way to long for to make thier stuff firewire. Disk dirve's usb, most of our burners are usb, palm connector, serial adaptor.

The thing is, everything on that list except the burner is fine for USB. Usb (even USB2) isn't designed for long sustained data transfer (i.e. burning); it works in bursts. Keep the little stuff on USB, and get the better stuff on firewire.

One problem

[chainrust]

There is no "master", which limits speed. It should be somewhere between USB 1 and 2.

Use for storage

[chaidawg]

Too bad I cant update the electronics I already have. I would love to jack my camera into my Archos jukebox to store pictures. That would be more useful to me than plugging it into the printer. Where there is a printer there is a computer. But for a longish trip (2 wks-1 month) using the mp3 player for picture storage would be great.

Re:Yeah, but isn't bluetooth insecure?

[Zathrus]

Yes, and your data is also going to be "transmitted out to the world" over a null modem cable. And ethernet. And your monitor.

You ARE fully tempest shielded, right?

What a load of crap. If you're concerned about someone intercepting the data during transfer then it's up to you to encrypt it appropriately.

And, frankly, nobody could give a crap about your precious pr0n collection anyway. Far too many people are far too concerned about stuff that nobody else wants to see in the first place.

Where is USB going?

[mustangdavis]

If they keep trying to replace firewire .... then:

Ans: USB On-the-Go Go Go Go ..... Gone!

IDIOTS...

[Eric_Cartman_South_P]

Why not just make the connection wireless? Come up with a standard that has short range wireless. Call it BlueTooth or something stupid like that.

Oh... wait...

I'm sick of wires. It's the fscking year 2002 and everything has wires pouring out of it. Apart from electricity, I don't want any wires anyware on the outside to sync, hook-up, etc.

Re:IDIOTS...

[Locutus]

That's right but remember how Java worked on Windows? Microsoft didn't want REAL Java to work so they hacked there own to keep Java from spreading. Microsoft does of 90% of the desktop market after all. Well, they don't like Bluetooth either and most of the FUD about Bluetooth over the last 2 years was most likely funded by that company in Redmond. Devices are popping up with WinCE on it and that's bad but when you add the fact that these devices could get more useful with a lightweight wireless system that spells trouble for WinCE and that means trouble for spreading/controlling the application APIs owned and wielded by Microsoft.

Why do you think they are now doing 802.11 hardware? Very soon they will release Win-Fi and pay hardware manufacturers to include it in the MoBo over Bluetooth....

Wires are silly alright but innovation is constantly throttled by very large companies with an agenda for protecting their monopolies. Be it Microsoft or Intel. USB will die soon but not as fast as we like.... Vote with your wallet and purchase Bluetooth product ASAP. IMHO.

LoB

1394 already does this

[Mr.roboto]

it's only slightly slower then USB 2, and it's cheap as can be. I need to get a card for my desktops so I can network em all, try here for a benchmark:
http://www.homenethelp.com/web/review/ Firewire-InH ome-Networks.asp

networking is easier to get cabling for as well, becuse of the wide avalibilty of double ended cables as a standard item.

Re:Jini

[elmegil]

I had exactly the same question. JINI was touted as the ability for devices to self-register and interoperate. I suspect it was intended to be medium agnostic (USB/Firewire/bluetooth), but why don't we see any implementations?

Still Not There

[Salamander]

USB OTG is still not really symmetric. It's just a way for devices to negotiate over who gets to be master; that master then takes over all the polling that the computer would be doing in traditional USB. It's still a fundamentally crappy way to do things, it wastes resources (which the consumer does pay for), and it only works for two devices instead of N. Firewire is still way better technically, and here today.

it's wired! So it requires a common HW interface..

[Locutus]

Sorry but wireless is the way to go for what is projected as the use for the USB-to-go. Think about it, how many PDAs and handheld devices have a common connector? That's one very large pile to have to move all in one direction and I don't think it's going to happen.

This is a pretty obvious problem and so I would think that this bit of PR is more of another way to stall Bluetooth. Find out who is behind this and not just the front-man and you'll probably find a hidden agenda.

Seriously, this is a stupid idea given the realities of how devices are connected today. IMHO.

LoB

Misuse of terminology - not exciting technology

[ngoy]

I fail to see what the effect of this will be. In reality, there is no such thing as a "peer-to-peer" network. Regardless of what communication device you are using, one of them HAS to ask for SOMETHING. That is the whole basis of communication. Question and answer. This role may be reversed many times during the communication, but it still exists. So you don't need a pc. Wow. A limitation born out of the technology they decided to implement it with.

Maybe someone will come up with the telephone next.

Shango

power slave

[shibut]

The master-slave now is not just for drivers but also for power. Under the USB gogo you'd still have to specify a device that will supply the power to the other USB device (one of nice features of USB is that the devices don't need to be plugged into a power supply) so connecting 2 "power slaves" to transfer files would not work.

Master and Servant

[_Sambo]

USB on the go is not Peer to Peer. The cable connecting the devices will determine which device is the Host (read Master) and which is the Peripheral (read Slave). They've also got two new connecting protocols. Each device must be a Dual Role Device. (DRD is my new TLA for the day). I understood the documentation on USB On the Go to say that each device that is compliant will have the drivers of the other devices that it will work with. Does this mean that they will be severely limited in what devices they will work with. (e.g. only HP cameras will work with HP printers, PDA's etc.) I found more info at: http://www.usb.org/developers/onthego/ The PDF presentations regarding On-The-Go are somewhat annoyingly colorful, but they may be trying to yak in marketingspeak.

I'm offended!!

[goldspider]

"Apparently this new technology is an offshoot of USB2 and it can remove the limitation of the master slave operation of normal USB devices, where you need a Host PC (the Master) to talk with the peripherals (the slaves)."

Comon people, this is the 21st century for Christ's sake! As progressive, technologically savvy people, we should be horrified that the institution of slavery is still a common practice in the computer industry.

On a more serious note, I wonder when the PC (politically-correct, not personal computer) crowd will take exception to the technological "master/slave" terminology.

Re:I'm offended!!

[DLWormwood]

I wonder when the PC (politically-correct, not personal computer) crowd will take exception to the technological "master/slave" terminology.

Around about the time they finally wise up to our use of "male" and "female" for connector cables and plugs...

Re:it's wired! So it requires a common HW interfac

[aron_wallaker]

I would think that this bit of PR is more of another way to stall Bluetooth

No one has to do anything to 'stall' Bluetooth - they've done that all on their own.

(Speaking as someone who was working on a Bluetooth project 4 years ago and still hasn't seen anything decent come to market)

But at least I'm not bitter.....:)

Re:USBUSB networking

[cheezedawg]

In order to network computers together with USB, you need to go through a USB bridge (which can support more than 2 computers). If you plugged a USB cable directly between 2 computers, it would short out the power supply and likely damage the system. That is why the type-A to type-A cables are illegal in the USB spec.

A lot of my stuff is domotic...

[Steve Franklin]

Except the OS, which is demonic.

The big hurdle, I would surmise, is figuring out a good reason for attaching your toaster to your digital camera. Now if the camera had AI, it could look at the toaster and then send an infrared signal to your smart alarm clock that would synthesize a real loud Kachung! sound like the old toasters make and give you an incentive for getting up.

Your bathroom scale could be linked to the refrigerator door lock to help you lose weight.

Your reading lamp could be linked to the kid's stereo--one one, one off.

Seriously, it would be nice to hook things together easily, but as someone already pointed out, you still need drivers. Of course you turn the mobile OS in say your PDA into bloatware like Windows by including every driver under the sun, it might work, after somebody invents the super battery. Maybe just add a USB port to all electrical outlets. That way you could just plug the hardware into the main power supply.

There is such a thing as taking modular too far.

Re:Hard drives

[Yohahn]

It already meets it, and because there aren't incumbering patents, it is cheaper to implement.

OK People, Pay attention!

[handorf]

USB:
Low speed peripherals (Keyboards, mice)
Low price peripherals, medium bandwidth (scanners, CDRW, small hard drives, mp3 players)
Firmly entrenched, all new PCs have USB 1.1 at least
Cheaper to implement.

Firewire:
High speed devices (Hard drives, video cameras, etc)
More expensive to implement
NOT FIRMLY ENTRENCHED!

USB is here to stay, people. A Firewire mouse just isn't going to happen. A Firewire scanner is a waste of $25 to implement the firewire on the scanner and the motherboard to support it.

Please stop with the "Who cares? Firewire is better!" If you have a PDA with a firewire chip on it, I'd like to see it! (A real PDA, not a very small PC).

This does matter, if you don't care, go back to the "Why buy a Toyota? An F-18 is faster!" threads.

Re:OK People, Pay attention!

[tuffy]

That's great that all you college kids have the money to buy new computers, but for us in the real world, using computers *gasp* more than a year old *gasp*, serial & parallel and the PS2 ports work just fine. They're cheap, and they always work.

Serial, parallel and PS/2 ports are nice, as long as you don't have a lot of peripherals and plan on using the IBM-PC platform forever. Personally, I was glad to switch over all my crap to USB if only to avoid juggling four different types of connectors (if you count the joystick port) that all need to be hooked onto the back of the box rather than a nice accessible hub.

USB may not be a great standard. But it is a standard, it's here now, and it works on Linux.

Re:OK People, Pay attention!

[rakerman]

Firewire is also the interface for the new pro digital cameras (10+ megapixels). A FireWire card only costs like 50 bucks. It's no big deal.

Bah

[Cinematique]

Call me crazy, but I've never seen the point in dueling standards. Shouldn't the whole point of a standard to get everyone in the same industry involved in its creation, and in the end, simplify a problem?

one standard = harmony
two (or more) standards = proprietary mess

USB v. USB2 v. USB-LMNOP v. FireWire
NTSC v. PAL v. SECAM
Beta v. VHS
DVD v. DVD-R v. DVD-RW v. DVD+RW v. DVD-LMNOP v. DVD+LMNOP
OpenGL v. DirectX v. Glide
Java [Microsoft] v. Java [Sun Microsystems]

and the one that really ticks me off...

HTML/CSS [IE Windows] v. HTML/CSS [IE Macintosh] v. HTML/CSS [Mozilla/Netscape] v. HTML/CSS [Everyone else]

bla bla bla bla bla...

ENOUGH ALREADY.

Re:it's wired! So it requires a common HW interfac

[Locutus]

So YOUR Bluetooth project never worked? I've not designed anything with Bluetooth but it was pretty obvious Microsoft didn't want it to work. Because there aren't many products out today isn't always an indication of it's abilities. I'll be Intel paid out a pretty penny to get USB on all those system boards out there and to keep Firewire off those boards. Not to unlike how Microsoft paid ISPs to not use Netscapes browser and gave them theirs for free at a time when Netscape had the better product.

So, can you tell us more about YOUR Bluetooth project?

Re:USBUSB networking

[cheezedawg]

I wasnt trying to say that it can't be done. The "specialized cables" that you were talking about just put a bridge in the middle of the cable, but there are bridges that can connect more than 2 computers. I was just saying that you can already connect more than 2 computers even without USB OTG.

http://www.linux-usb.org/usbnet/

Re:it's wired! So it requires a common HW interfac

[aron_wallaker]

The biggest problem with Bluetooth has never been whether 'a' project works, it's been whether umpteenth vendors can all write their own software stacks of a very complicated (and, back then, very "1.0") communications protocol and then all show up at Comdex expecting their widgets to talk to each other. I saw that problem coming from miles back - the Bluetooth 1.0 intervendor 'interop' was nothing short of a disaster and it shouldn't have surprised anyone. Now it's almost 2003, shipping Bluetooth implementations are out there...but they're relatively rare and the average person on the street has never heard of it, doesn't know what it is and exerts absolutely zero market pressure for it. They created an incredibly complicated solution for a simple problem (cable replacement) which, it turns out, few people want. It reminds me sooooo much of IrDA...which I also worked with. (My career in device drivers was a little checkered, to put it nicely.)

As for Intel 'paying' to not have Firewire on the motherboards.....who were they paying ? At the time of USB1 vs 1394 the majority of motherboards were using Intel chipsets. You can accuse them of being biased against firewire, but it didn't take any payola - it was their market to steer. We have more competition now, but most of the other motherboard chipset vendors are trying to undercut Intel on price, they're not going to worry about Firewire.

Re:Hard drives

[Yohahn]

You are full of FUD.

How do you support your claim that audio/video is standardized on firewire? The DV cam is the only bit that is very prevalent. These things can change with 1 generation of purchasing.

You haven't supported your claim about the protocol speed either. Show me the numbers.

I think there's a foot in your mouth

[Johannes]

There's 3 main host controllers for USB: OHCI (1.1), UHCI (1.1) and EHCI (2.0). There are some other HC's, but they are solely for embedded systems and rare.

ALL three of those are Interrupt (IRQ) driven.
ALL three of those have a schedule which is INDEPENDENTELY executed.

There is NEVER busy waiting in USB drivers during normal operation (some HC's will require a busy wait on a register to reset it when you first startup the system, big deal).

It's amazing how wrong you are.

Apples and Oranges

[Andy Dodd]

If USB-to-go-go-go-go-whatever were based on USB 1.1, you'd have a valid point.

Comparing USB 1.1 and Firewire is like comparing apples and oranges. They have totally different targets, costs, etc.

But USB-to-go-whatever is based on USB 2.0, which is no cheaper than Firewire. Therefore, why not go with Firewire, which is more mature?

USB 2.0 is going to have a long fight ahead of it. Firewire is a better protocol, is more mature, and firmly entrenched in the video market, which is one of the "killer apps" for high-speed serial. And it isn't any more expensive.

USB 1.1 is here to stay and will be around for a long time. For low-bandwidth devices, it's perfect.

Re:Apples and Oranges

[handorf]

I won't disagree with your points on Firewire being better/more mature/etc than USB 2.0.

Personally I haven't been overly impressed with USB 2.0 myself.

But given that USB 2.0 is backward compatible w/USB 1.X, putting a USB 2.0 choice on a PDA and then implementing USB-to-Go-Go-Gadget-Interconnectivity will probably wind up being the choice most manufacturers implement. It lets all of us who have no firewire OR USB 2.0 ports still communicate with the latest & greatest PDA and still get kewl features like inter-device-connectivity if we happen to have more than one USB 2.0 device.

Re:Careful

[Yohahn]

so, ok, grep for irq.

I believe you'll still find a interrupt for things like short packets and the like.

Great...

[mmol_6453]

Just when I thought that the USB architecture had it right.

First, a story about how a new biometric device is more secure...except that it could be disconnected from the computer. Then there were several comments about the ease of access to the PS/2 port's hardware address, for trojans to sniff. (1, 2, 3)

Now USB is looking at going P2P. That's not a good idea, since even switched networks can be confused by ARP cache poisoning. (Which there surely would be an analog to in any switched P2P network)

Re:OT: Distributed computing

[Saeger]

Drug patents have been assumed by governments before

That's the main reason I continue running the UD client (on all my 'puters). I know that even if they wanted to hold the cancer cure/treatment for ransom, the world (minus pharma shareholders) wouldn't let them get away with it.

--

Close, but not quite

[SWPadnos]

Actually, USB on-the-go isn't really a part of USB 2. It's a second layer protocol that allows devices to act as either a master or a slave, depending on the task at hand. Although the on-the-go spec is an addendum to USB 2, there is nothing in it that requires 2.0 functionality to work.

An example: a digital camera that can act as a master when connected to a printer (pretending to be a computer), and as a slave when connected to a computer (pretending to be a storage device).

There is still a master and a slave, and the cable determines which unit starts out as the master. This is done with a new type of USB connector - the mini-AB. (the old ones were type A, the flat ones, and type B, the square-ish ones).

The Mini-AB jack can accept either a mini-A or a mini-B plug. The device that the mini-A is plugged into acts as master, and the device that gets the mini-B is the slave. (The protocol allows the master to pass control to another device) There are other cables for connecting these devices to "legacy" USB ports - mini-AB to A or B cables. These cables are wired so that the OTG device knows whether it should be master or slave.

As for the devices only working with one manufacturer's peripherals (someone mentioned an HP camera only printing to an HP printer), that may happen. Although the "class" drivers are more likely to be implemented in embedded devices, there are probably features that won't work when mixing and matching devices from different manufacturers.

strikes me as pointless

[g4dget]

I don't see this as a big deal. It's easy to create a little "active cable" that has all the intelligence necessary for, say, a digital camera to talk to all the devices it might want to talk to. And by putting the intelligence into the cable, I can upgrade it without having to get a new camera. In particular, if my camera doesn't work with my printer because of some quirk or new feature in my printer, I can just buy a new cable.

Or, if you do want to put the intelligence into the devices, you could create a device-to-device networking cable, analogous to the host-to-host networking cables, allowing any device to talk to any other without any changes to its hardware (the software, of course, would need to be upgraded).

I have never seen this as a big advantage of Firewire and actually think the feature would be best left out of any standard. It's more important to get standard profiles for things like serial ports and other devices so that people can figure out how to build gadgets that do connect them.

Might have been posted already

[strictnein]

but plenty of digital cameras and printers already support communicating via the USB port, without a computer (HP DJ7150,7350,7550,P130, etc).

whatever though... i'm sure this standard is better for that sort of thing...

For the thousandth time, repeat after me:

[Catbeller]

but it is sans the license fee to Apple

I hear this canard so many times I wonder if Intel assigns munchkins to spread it on message boards.

For the thousandth time, repeat after me:

Apple does not charge a licensing fee for Firewire.

Didn't know that...

[Eric_Cartman_South_P]

most of the FUD about Bluetooth over the last 2 years was most likely funded by that company in Redmond

I didn't know that! Having worked for a Java developer for 4 years (doing something different now) your example of what M$ did to Java made it crystal clear, and now a lot of things I read make sence. MS keeps Bluetooth down and their own crap comes to power. Look at Sony and their new palm 5 device. has a Comfact Flash slot except for the fact that its HACKED so only a sony wireless WiFi product will work. No other CF mini-hard-drives or any other product. Heck, with the build in video camera, and a 3rd party wireless phone/data CF card, you could have a hand held VIDEO PHONE in a few weeks! But because of SONY and their restriction to, as you said, protect their monoloplies, we'll have to wait until 2020 for such a prodcut.

Now I feel sad. Maybe I'll d/l some .ogg music :)

Re:Didn't know that...

[Locutus]

When you see ANY news related to Microsoft, just ask yourself, "How does this protect the Windows monopoly?" and you'll recognize what's up. It helps seeing the bigger picture in the market like who is doing what in game consoles and who's doing what in the PDA sector.

BTW:
I just read that Microsoft is lobbying Washington for more broadband for the masses. They don't do squat for anything but to protect Windows so this must be about getting their bloated systems connected in the homes. The pressure is on them to do this quickly before someone else's standard picks up and beats 802.11 into the home market.

They really need broadband or nobody will want their DRM based OS's because there won't be much to protect. Surely not enough to sustain the whole company.....

I'm all for more broadband but if it means Microsoft fails again.... go 56K dialup! ;)

LoB

Re:it's wired! So it requires a common HW interfac

[Locutus]

That's too bad. Sounds like there needed to be a few Bluetooth compatibility labs so people could play with other in a lab and figure what's going on. Kinda like IBM's labs they setup for things like Linux, Java, etc. too often standards bodies leave too many doors open because of commitee member pressure and the standard has to hash it out on the street for a few years. Companies like Palm, Handspring, etc sould have replaced IRDA with Bluetooth years ago since they only needed to work with themselves anyway.....

Regarding Intel, you are probably right. Intel killed the chipset competition with the SLOT-A CPU and took the MoBo market to themselves for a number of years. Many had to license their chips just to stay in business because of the proprietary slot interface. Probably why USB came out then since Intel is used to planning 5+ years out.
IMHO

LoB

Print Directly From Camera

[Pooua]

So using this, theoretically you can print using your digital camera directly on your printer

Couldn't the camera simply act as the host?

My USB-base Olympus C-3040 already allows for this (though I've never used it).

Re:OT: Distributed computing

[Abreu]

Sorry to be pessimistic, but it seems that you forget the golden rule: He who's got the gold, makes the rules

I dont think anybody (short of a goverment, such as South Africa did) would be able to fling the finger at Big Pharma