Which external connection has the highest data throughput with respect to the headers used? I would appreciate a slowest to fastest list including USB 2.0, USB 3.0, FireWire 400, FireWire 800, eSATA and Thunderbolt considering bandwidth that is consumed by command and control (reducing the data bandwidth availability).
The theoretical maximums are as follows:
In bits per second, that is:
- USB 1.1 = 12 Mbit/s
- Firefire 400 = 400 Mbit/s
- USB 2.0 = 480 Mbit/s
- FireWire 800 = 800 Mbit/s
- USB 3.0 = 5 Gbit/s
- USB 3.1 = 10 Gbit/s
- eSATA = Up to 6 Gbit/s (750 MB/s) right now as it depend on the internal SATA chip.
- Thunderbolt = 10 Gbit/s × 2 (2 channels)
- Thunderbolt 2 = 20 Gbit/s
- Thunderbolt 3 = 40 Gbit/s
In Bytes per second, that is:
- USB 1.1 = 1.5 MB/s
- Firefire 400 = 50 MB/s
- USB 2.0 = 60 MB/s
- FireWire 800 = 100 MB/s
- USB 3.0 = 625 MB/s
- USB 3.1 = 1.21 GB/s
- eSATA = 750 MB/s
- Thunderbolt = 1.25 GB/s × 2 (2 channels)
- Thunderbolt 2 = 2.5 GB/s
- Thunderbolt 3 = 5 GB/s
However, this does not provide the actual answer. As an example, FireWire 400 is a serial connection. The entire 400 Mbps is available for data transfer. USB 2.0 sends command and control data through the same connection the data uses limiting the 480 Mbps connection to 380 to 400 Mbps. When considering throughput the list looks entirely different.
For the speed/throughput/bandwidth of more devices have look at this article on wikipedia http://en.wikipedia.org/wiki/List_of_device_bit_rates#Peripheral
Wikipedia gives a quite comprehensive comparison: http://en.wikipedia.org/wiki/Serial_ATA#Comparison_with_other_buses
Intel Thunderbolt, as per the Wikipedia SATA link just above, is 10Gbit/s.
Also, none of these answers so far give any practical/useful information. Theoretical maximum and real-world speeds can vary wildly, and only some significant actual testing will give meaningful answers.
So far, I haven't found many such tests. There's one at Crunchgear.com:
But even this leaves some question, as perhaps their specific USB 3.0 implementation is not optimal. (we need more variety to be sure, and even then, your system (or any given system) may not produce comparable benchmarks)
Another seems to suggest USB 3.0 "Turbo" (whatever that is?) has a bit over eSATA, at sansdigital.com:
But I have to question that, suggesting ~200 MB/s hard drive read/write speeds - unless hard drives have dramatically improved recently, I don't believe those speeds are physically possible, and suspect those speeds are just cached.
It's probably relatively safe to go with eSATA or USB 3.0 and get speeds that are close to optimal... as long as there's nothing choking your chain, so to speak. (poorly designed or cheap component, etc., causing a bottleneck) We really need more real world comparisons with various different hardware components.
This is a perfect example of an incredibly poorly worded question arriving at the incorrect "popular" answer.
"Fast"??? What does that mean?
I have 80 people that I need to move one mile as fast as possible. My choices for vehicle are:
a School bus, top speed 70 miles per hour
a Lamborghini, that can do 220 miles per hour
So which do I choose? The "faster" one?
The point I am making is explained here...
USB 2.0 can push (theoretical maximum) 480 Mbps
Firewire 400 can push (theoretical maximum) 400 Mbps
HOWEVER, the USB bus charges you approximately 20% overhead. This means that the theoretical maximum throughput data rate is closer to 384 Mbps. This is because 20% is used for controlling the bus. That is, control signals are sent through the same pipe that is used to move data.
Firewire does not have this restriction to consider. So in data throughput FireWire 400 will beat USB 2.0, even though USB 2.0 has a higher theoretical maximum.
This is not the only example of why answering a question like this by citing theoretical maximums does not provide a correct answer.