Conclusion: Multi-TT Worthwhile For High Loads
Admittedly, with a test set-up as extensive as this one, you have to take some time to look over the benchmark scores and let them soak in. One thing you'll notice right off the bat is that the more USB devices you have, the lower the overall performance is. For example, the read performance for the external SPIO drive (USB 2.0) drops from over 23 to some 20 MB/s when you have a webcam and a USB 1.1 Memory Stick running. The drop is even more pronounced in the write performance.
It's much the same story, albeit with a more dramatic ending, if you expect good performance from a USB 1.1 device such as a Memory Stick. Once the webcam starts uploading images and the SPIO begins churning through data via USB 2.0, you have only a fraction of the original 900 kB/s left over.
And that is exactly where LinXcel's multi-TT hub shines. If you're more interested in the SPIO, which is the 2.0 device, the differences at high USB transfer rates are insignificant. That applies especially to cases where there is only a small sliver of bandwidth left over from the already paltry transfer capacity of a 1.1 device.
You'll be confronted with this scenario whenever you run several USB 1.1 devices on one 2.0 hub and/or at least one of these devices uses the isochronous protocol of the USB bus. That definitely spells a drop in performance.
If you're using several USB devices without a hub, make sure you find a multi-TT device. The performance boost will be modest in a pure 2.0 environment and enormous if you have many 1.1 devices running. You might, however, have some problems identifying a multi-TT hub because the packages only rarely mention this feature at all.
