There has been a lot of talk in the past few weeks about the new serial interconnect, Thunderbolt, developed by Intel and Apple. But Subu Sankaran of Future Devices Technology International has his reservations.
The newly released Thunderbolt interconnect provides two channels of bi-directional data transfer at 10-Gbps rates, which admittedly is way beyond an order of magnitude more than what the USB 2.0 protocol offers. In the future, that will be bolstered even further, as the interconnect will incorporate optical (hence its original name of "LightPeak") as well as copper cabling. However, this is still likely to be some way off.
The huge jump in bandwidth Thunderbolt represents means that, in principal, it should prove to be a very attractive prospect. Despite this, there are serious technical and economic drawbacks that need to be reflected on. All this added performance does not come cheap, and so its ability to swell a bill of materials might actually overshadow its ability to transfer data.
The added expense of implementing Thunderbolt should be relatively easy to cope with in larger items such as computers, but for the pieces of peripheral equipment (such as cameras, storage devices, etc.) having an interface chip that supports it will raise their price tag considerably –something that will be tricky in such commoditized areas.
Though Thunderbolt is compatible with the DisplayPort interface, opening up possibilities for it in application areas such as digital signage, in nearly every other non-consumer market USB is deeply entrenched. There will be difficulties applying the new interconnect to existing hardware and the cost involved will be hard to justify as the benefits of higher speed connectivity won't be as apparent in these environments. Few engineering teams here will take the risk of choosing it if there are still doubts about how great its proliferation will be.
USB, by contrast offers a far more practical migration path, with new iterations increasing the speeds that can be sustained and the power efficiency. Furthermore, it has billions of ports already in operation.
Some people may have prematurely chosen to imply that Thunderbolt represents the end of USB, but in truth there is little to back up this conclusion. More likely is that this interconnect, much like FireWire before it, will remain a nice addition to Apple's Macbooks, iPads and the like, allowing the company to differentiate itself from the competition, rather than seeing widespread adoption across either the consumer or the industrial space.
By compromising on the original concept of moving to an all optical mechanism (as there simply wasn't the demand for the bandwidth this would support and the price points were just too high) Thunderbolt positions itself too close to USB's home turf. It is likely, therefore, that in a similar fashion to FireWire, this will end up being categorized as little more than niche—feeding certain OEMs vanity rather than customers' real requirements. Conversely, USB is a highly cost-effective, backward compatible connectivity solution which has already been embraced universally. As a result it can rely on a much greater groundswell, and will thus remain the way by which the majority of the world's hardware interacts for a long time to come.
Subu Sankaran is an account technical manager at Future Devices Technology International, a company that specializes in converting peripherals to USB. Sankaran is a semiconductor industry veteran who formerly worked at ST-Ericsson and NXP Semiconductors and holds an MBA from the University of Phoenix and a bachelor of engineering degree in electronics and communications from Bangalore University.