Firewire (IEEE1394) slowly vapourising?

[DrGeoff]

I was hunting around for a microATX motherboard to house in a rack case for part of a mobile recording rig, but have been unable to locate any newer motherboards with firewire ports onboard. Firewire is necessary since my interfaces require it.
I noticed that laptops also seems to have removed the firewire port as well.
About the best I could find was the Gigabyte module, which is LGA775 and a bit old now (also hard to come by since it is discontinued).
My preference was to use a motherboard with the TI chip onboard rather than try and accommodate slots and card guides for PCIe cards. This was to be a recording rig only, no plugins or complex processing of the audio streams. 16 channels via the firewire interfaces (24/44K).
Any ideas?

[Uncle Vernon]

I was hunting around for a microATX motherboard to house in a rack case for part of a mobile recording rig, but have been unable to locate any newer motherboards with firewire ports onboard. Firewire is necessary since my interfaces require it.
I noticed that laptops also seems to have removed the firewire port as well.
About the best I could find was the Gigabyte module, which is LGA775 and a bit old now (also hard to come by since it is discontinued).
My preference was to use a motherboard with the TI chip onboard rather than try and accommodate slots and card guides for PCIe cards. This was to be a recording rig only, no plugins or complex processing of the audio streams. 16 channels via the firewire interfaces (24/44K).
Any ideas?

The reality is Firewire is heading the way of the Zip Drive and Betamax.  Plug-Ins are likely to be your only practical option and those will soon be harder to find.

[DrGeoff]

It's certainly looking that way. I suspect that the DV cam market, which used to have firewire for DV and is now card based, has helped to kill off the 1394 port. And now that Thunderbolt is starting to appear there's no need for the legacy technology. Unfortunately I'm not about to upgrade all my audio interfaces to thunderbolt right now, and will probably end us using a bridge adaptor for some time in the future (until I can afford a couple of Apollo interfaces maybe).
I was hoping to be able to put something together fairly quickly and economically for the short term. I did look at the MAC mini, but the DAW software runs like a dog on the MAC OS and I'd prefer to stick with Win7 if possible.
Maybe a PCIe card is the only way.

[David Aurora]

It's a shame that it never really seemed to catch on properly. I've been using Firewire a lot since 2001 I think, and it's flogged the zillion other connections around the studio and been solid as a rock. But it's always been a challenge to find stuff compared to shitty USB gear

[SeanB]

Saw a lot of sale of firewire PCI and PCIE cards today, each going for around $12 retail. I think I should get one of these to get the video of my DV camera before the camera dies totally ( Sony.....) from capacitor plague.

[DrGeoff]

Finally found one that takes a more modern processor (LGA1155) and has a couple of 1394 ports.
Intel DH67GDB3

[amspire]

My understanding is that Firewire has to go. It works by giving the attached device something similar to DMA access to the memory, and there are major security holes with this that cannot be fixed.

Protocols like USB and the newer protocols adds a  software later between the CPU memory and the peripheral, and so it is possible to fix security flaws.

Richard.

[DrGeoff]

My understanding is that Firewire has to go. It works by giving the attached device something similar to DMA access to the memory, and there are major security holes with this that cannot be fixed.

Protocols like USB and the newer protocols adds a  software later between the CPU memory and the peripheral, and so it is possible to fix security flaws.

Richard.

Supposed security flaws like this are largely irrelevant to someone who has a rack full of high performance interfaces.
I'd be more inclined to consider it a dwindling consumer interface, largely due to the evolution of dv cams moving to SDHC cards.
And now we have Thunderbolt, which is a 10Gbps serial interface appearing to take over the role of firewire. Have you heard of any security considerations with thunderbolt?
Unfortunately USB has never been much use for serious audio or video work, probably because of that software layer.

[Lukas]

Have you heard of any security considerations with thunderbolt?

Since Thunderbolt has DMA capability, there are the same issues as with Firewire. See http://en.wikipedia.org/wiki/Thunderbolt_%28interface%29#Security

[Monkeh]

I don't see anyone harping on about Expresscard having security issues. It has a direct PCI-E interface, too.

Let's not forget, people: If you have physical access to a computer, any and all security must be assumed compromised.

USB2 was useless for latency or transferring large amounts of data. USB3 is much, much better for throughput, but it still sucks for latency.

[DrGeoff]

USB2 was useless for latency or transferring large amounts of data. USB3 is much, much better for throughput, but it still sucks for latency.

Yes, I think the security issues are a red herring here.
Apple/Intel developed thunderbolt as a high speed replacement for 1394. It's a neat way to extend your PCIe bus outside the box and the performance is more than 10 times better than 1394b (FW800). UA released their interface (Apollo) with an option to plug in a module to support thunderbolt, for very low latency.

USB has never been satisfactory for audio interfaces intended for recording purposes. Attempting to stream 16 or 24 channels, even at 24/44K with low latency was impossible. It can be done with FW quite easily, however FW had issues with chipset compatibility and a lot was written about design flaws in some chipsets and motherboards. The TI chipset is the respected one for reliable audio work, and I've found the Via chipset also works very well.

Anyhow, the Intel microATX motherboard should fit into a 3RU rack box with an i3 CPU plus heatsink and fan. I'll add an ATX PSU and a couple of 2.5" SATA drives and should have a reasonable rig to rack up into the mobile unit with a couple of 896 interface boxes. I thought I'd position the motherboard to have all the rear panel connectors out the front of the rack box, for easy access.

[amspire]

I don't think security concerns are a red herring. I think the computer industry is starting to realize that consumers cannot be expected to know who to trust, and it is up to the computer industry to start building in genuine security layers. Apple is a company that has only started to get serious about security in the last year or so - before that, they were able to get away with the phony pretense that their computers were "more secure" just due to the fact they were not targeted.

Things like expresscard and Firewire came from a pre-security era - security was hardly considered at the time.

I am not convinced that Thunderbolt will succeed. I did a search for a Thunderbolt PCI/PCIe card on ebay and couldn't see anything - not a good sign. Is Intel licensing the rights to AMD, Marvell or anyone else to make chipsets or is this an Intel-only technology? If it is Intel-only, it will have a short life until a USB 4.0 or a new generation fast Ethernet comes along.

Intel may love the idea for a serial bus expansion interface, but I think the computer and operating system industry will realize they cannot go that way, no matter how much techies like us like the idea.

The big change is peripherals are now running full-blown linux-based operating systems full of third party software stacks. Even if you trust Canon, Nikon, Sony, etc, you do not know who wrote 90% of the software running on their devices. In many cases, the manufacturer probably does not know either. Companies like Sony have shown many times that they are clueless about security.

There will be a need for an interface that can stream real time 1080P and 4K video from consumer devices. I don't think the long term interface exists yet, but it will come. Expresscard, Firewire and Thunderbolt will all disappear into forgotten history. We will still have RS232 though and USB.

In the mean time, the Pro Audio and Video industry will still probably have to offer dedicated PC cards as an option for maximum multi-channel low-latency performance to their audio/video capture/streaming boxes.

Richard.

[Mechatrommer]

i have firewire port built in my intel motherboard (bought 5 years back) to stream (download) realtime uncompressed video from cassette to pc. with availability of "cassette'd" high end device (beta or videocam) still out there, i think firewire is still needed. + i dont thing current usb BW is capable of streaming realtime uncompressed video. i'm maybe outdated though, any suggestion welcomed.

[Monkeh]

Yes, because Expresscard comes from the ancient days of 2003, only eight years after IEEE1394..

The security concerns are crap. Modern Intel CPUs have an IOMMU, AMD CPUs should be catching up with that soon enough. The problem no longer exists if people bother to implement some security.

[amvakar]

Having used a cheap Chinese USB-to-RS232 adapter before, I can also say with confidence that there is nothing stopping a device from misbehaving and bringing down a system with the help of buggy drivers. I'm sure a more malicious enterprise could do much worse. Having said that, USB 3 and future revisions will own the consumer markets. Consumers do not care enough about real-time performance to seek out an expensive interface that solves no real problem other than the lack of $50 cables for every device they need to plug into their computer. Professionals will do differently, but their preference for Firewire hasn't stopped this thread from appearing.

[rsjsouza]

I think IEEE1394 will not disappear but instead be relegated to niche applications (just like DAT tapes and the SCSI interface), as it is the only true isochronous interface...

Regarding security, I tend to agree with Monkeh and DrGeoff - security is incredibly weaker in the social aspects (PEBKAC) than the peripheral interfaces.

If someone is really interesting in invading a PC via such a sophisticate way (read about Stuxnet, Duqu), just think about the thousands of USB-based products from obscure manufacturers around the world. Nowadays several "el-cheapo" microcontrollers have USB interfaces that can be set up to trap your system not via the interface itself, but by asking the user to install the "proper" device drivers from the "manufacturer". And we know that users don't give a "flurry animal's behind" for signed drivers - they click away for anything.

I can only guess how many USB-based products are plugged to PCs worldwide and people don't even think about security... That and the fact the IEEE1394 market is shrinking (similar to the MacOS/Windows virus dilemma) makes me more comfortable to plug a IEEE1394 device to my PC.

[nitro2k01]

If someone is really interesting in invading a PC via such a sophisticate way (read about Stuxnet, Duqu), just think about the thousands of USB-based products from obscure manufacturers around the world. Nowadays several "el-cheapo" microcontrollers have USB interfaces that can be set up to trap your system not via the interface itself, but by asking the user to install the "proper" device drivers from the "manufacturer". And we know that users don't give a "flurry animal's behind" for signed drivers - they click away for anything.

That's a different attack class all in all. For what you're describing, it would typically be adequate to implant some custom malware that won't be picked up by scanners, which either just live in userland, trick the user to provide admin privileges, or ping a server until the next 0day kernel exploit arrives. As for signed drivers, at least on 64-bit Windows, they can't be installed without a certain amount of hassle. Enough to discourage the average moron.

The Firewire attacks are unique in that you can operate in vivo on a running system. The only two times when you specifically need this kind of access is 1) if you need to pwn the system within 10 seconds 2) if you need to capture the running RAM. Say you arrive at a crime scene of a paranoid child porn sharer who runs everything in RAM or uses full disk encryption, so you absolutely need tp caåture the existing RAM.

However, to my knowledge, this attack has pretty much been closed down anyway, or at least can be closed down. (IOMMU, or making the driver not enable DMA mode.) I wouldn't discourage using FW for security reasons. USB can also be used for pwnage, for example through a bootable USB stick, or yet-to-be-explored attacks involving emulating a USB device that sends invalid data. Case in point, PS3 could be hacked with a TI calculator.

It was USB that killed Firewire by relegating FW to only the areas where the lower latency is actually needed, in practice DV cameras and high end audio interfaces.

[DrGeoff]

And then along came this.

http://www.wavemachinelabs.com/Products/auria

Could be an interesting mobile recording platform. I can't imagine that latency would be good though, since it uses USB2 for the audio capture interface.
 

[Uncle Vernon]

And then along came this.

http://www.wavemachinelabs.com/Products/auria

Could be an interesting mobile recording platform. I can't imagine that latency would be good though, since it uses USB2 for the audio capture interface.

USB on an Ipad, good luck, good luck achieving semi-pro audio with that.

[DrGeoff]

And then along came this.

http://www.wavemachinelabs.com/Products/auria

Could be an interesting mobile recording platform. I can't imagine that latency would be good though, since it uses USB2 for the audio capture interface.

USB on an Ipad, good luck, good luck achieving semi-pro audio with that.

I thought much the same. Imagine watching the ipad during a recording session suddenly decide to do some apple thing and stop.

The quality of audio and sample rates are determined by the mics, preamps and interfaces. The ipad is only a sample capture device  with a fancy gui. Wouldn't dare use it to mix or use any effects or eq through it.