Alex headshot

AlBlue’s Blog

Macs, Modularity and More

FireWire is dead; long live USB!

2005 Mac

It seems like FireWire (AKA IEEE 1394 or iLink) is dead, but just not buried yet. Once the darling of the Apple eye, its transition to USB 2.0 has already begun. One only has to look at the iPod system requirements:

iPod nano and the Fifth Generation iPod will charge via a 6-pin FireWire connection but will not sync over FireWire.

Documentation suggests that FireWire support was dropped from the 5th generation iPod (and nano) because there was a need to compress the iPod and there wasn't enough space to build in FireWire chips. I don't buy into that argument; most of the thinning down on the iPods has been because of the size of the hard drive inside rather than the circuitry. If a drive goes from 2 platters to 1 platter with an increase in the number of bits per platter, you can get the same storage with reduced space. Secondly, the Nano uses flash memory chips directly instead of being wrapped in a Compact Flash package, so the design can be shrunk even further.

I believe that FireWire was removed from the end user products as a first step towards eliminating FireWire from mainstream Macs. It's the main reason why Macs have FireWire in today's market (other than video devices such as iSight). It's no secret that Intel has always promoted USB over FireWire because USB requires a host OS to process incoming USB requests; as opposed to FireWire, which provides processing in the device itself. (It's also why FireWire tends to be marginally more expensive than USB, because USB devices are dumb.) With the upcoming switch to Intel processors, it's quite likely that an end goal of getting rid of FireWire in replacement of the next-gen USB (3.0?) is likely to result in new, high-end Macs getting cooler technology first.

So how come we are in a situation where FireWire is getting the boot, anyway? It boils down to three fairly simple reasons:

  1. A plethora of different cables. The fault of the specification, more than anything else. When FireWire was initially conceived, there were 2 connectors; a 4-pin and 6-pin cable. With USB, there's only ever been 1 cable. The idea was that some self-powered devices (e.g. video cameras, Sony Playstation etc.) didn't need the power provided, so only had a 4-pin arrangement (with two pairs of cables). Others that needed power would have a 6-pin (4-pin plus power/ground). For FireWire 800, the cable changed again; this time, adding a couple of reference ground signals (to reduce radio interference) as well as an additional unused status signal. Even though they're all backward compatible (you can plug a FireWire 800 device into a 6-pin or 4-pin FireWire 400 device), the fact that the cables look different means that consumers don't know this. Also:
  2. FireWire 800 never took off. This is Apple's fault, plain and simple. They bundled FireWire 800 (AKA IEEE 1394b) only on their pro-level Macs (PowerBook, PowerMac) and left all of the consumer level devices (iMac, iBook) with FireWire 400. As a result, there was never a large market for FireWire 800 devices created, so the devices that were created were expensive. Plus, USB 2.0 (with a transfer rate of 480mbps) trumped FireWire 400 (400mbps), so on a recent iBook, USB 2.0 is faster than the FireWire port on that machine.
  3. FireWire problems zapped data. A mixed fault of both Apple and Oxford Semiconductors, in which an Apple update managed to trash all data on an external FireWire device. Problems in upgrading aren't uncommon, but even to this day I am especially paranoid about upgrades with my external FireWire disk attached. (Helpful comments like 'remove it first' don't help ... I boot off my FireWire disk, so that's the only way to upgrade my system.) The exact blame doesn't lie with any one particular fault, and the problem wasn't widespread, but it certainly kicked technology in the nuts.

It's a real shame, because like the Power PC, FireWire was an Apple-involved specification that was both ahead of its time and undervalued. For example, not many people know it's possible to hook up two Macs with a FireWire cable and use IP-over-FireWire to network the two together. On a couple of FireWire 800 enabled Macs, you can actually get better transfer rates than over Gigabit ethernet (though the difference is minor). Also, Apple machines can be booted into a 'FireWire target disk' by holding down T whilst the computer boots; this turns it into a very expensive external hard drive. (Where it really comes into its own is when the disk is corrupt and needs verifying/repairing/reinstalling; you can use any external FireWire tools to manage the disk.

A couple of years ago, you could see the future plans for FireWire developing; there were to be FireWire 1600 and 3200 specs (1394c and 1394d respectively) which would have meant faster-than-gigabit networking. There were even plans to run FireWire 800+ over Cat-5 UTP cables (ethernet cabling) which would have made connecting systems together a snap. These are all stillborn now, and likely will never see the light of day.

The moral of the story is this; if you want to see a new high-end technology take off, don't just put it in your high-end systems. Make it available in the lower systems as well, and you drive the market; the market then ensures a plethora of devices, and your technology succeeds. Ironically, this is exactly what happened with USB when Apple moved away from the ADB mouse connections and replaced them with USB devices; the rest of the industry moved forwards and adopted USB as a defacto standard. Ironically, had Apple put FireWire 800 on all systems when it first came out, we might have FireWire 1600 announced at the Apple Expo in January. Nowadays, even the low-end systems have 802.11g and Bluetooth 2.0; it seems that Apple may have learnt its lesson the hard way.