Thunderbolt Sideswipes USB3.0 – or does it?
Apple introduced their new MacBook’s last week and like any new Apple product, there is always some interesting technical buzz to capture the imagination of the press and Apple’s loyal customers. Apple didn’t disappoint with their new laptop rollout that included the introduction of Thunderbolt.
The big question is – will Thunderbolt kill USB3.0 before it has even begun to become popular? The simple answer is no.
Sure, Thunderbolt is 10Gbps and can multiplex PCI-e and a DisplayPort on one nifty connector that is DisplayPort compatible. USB3.0 is poky by comparison; it is spec’d at 4.8Gbps, but practically speaking is more like 3.6Gbps and in real products shipped to date has been closer to 800Mbps of throughput. Further, USB3.0 does not include DisplayPort. So why isn’t Thunderbolt the coup de grace for USB3.0? Three reasons; Cost, USB compatibility, installed base.
Cost. Let’s face it, Apple’s customers aren’t cost conscience buyers. If you are a chip person you can look at the photo of the Thunderbolt controller chip and tell it’s expensive. More importantly, the peripheral product hanging on to the end of the Thunderbolt link requires its own Thunderbolt controller and then a unique chip to interface to the port of choice. That may be fine for a high end, high performance HDD, but for many other peripherals, not so much. Since the output of Thunderbolt is PCIe, the remote product essentially includes a PCIe to USB or eSATA or Whatever-your-choice controller, just as if it were on a motherboard. So for illustration, let’s take a USB hub.
Today, if you wanted to build a USB3.0 hub that can also support USB2.0 peripherals you slap a USB3.0 hub chip onto a board along with some power supply chips and you are done. If you wanted the same USB3.0 hub but wanted to connect it to Thunderbolt, you’d need a Thunderbolt controller, which is probably 3X the price of a USB 3.0 hub chip, plus a PCIe to USB controller chip and most likely a USB3.0 hub chip since most USB host controller chips only have two ports. In simple terms, the cost of a Thunderbolt to USB hub will be 4X to 5X the cost of a simple USB3.0 hub. Time and Moore’s law will change that, but it will take a while.
Now let’s look at USB compatibility. The new USB3.0 connectors are backward compatible and protocol compatible with the many billion’s of USB 2.0 products that are in the market today. You may have a new USB3.0 laptop and not even know that the ports are capable of USB3.0 speeds since they look just like the USB2.0 ports you already know and love. They just work much faster when connected to new USB3.0 capable peripherals.
Finally there is the issue of the installed base. Every year over 2.5 billion new products ship that have a USB port. There are over 10 billions products in use today that have at least one USB port. Every printer, cell phone, USB HDD, USB memory stick, keyboard, mouse, etc all have USB ports. It truly has become the Universal Serial Bus. Therefore, paying a 4X premium to have a Thunderbolt to USB adapter won’t make economic sense – unless you are an Apple customer.
Don’t get me wrong. I can get as geeky as the next guy and think that Thunderbolt is a nice new technology. I just don’t expect it to cut short USB3.0 – even if it is faster and has the DisplayPort cool factor. I do expect Apple enthusiasts to be able to buy some very nifty over priced peripheral products, and do expect that there will be some kick ass fast HDDs that will fully utilize the throughput of Thunderbolt. However for most products the very low price point and compatibility of the installed base of USB products favors USB3.0. That is precisely why Intel positioned Thunderbolt as a complimentary technology and not a replacement technology – because it isn’t.