After the demonstration of Willamette during the first few hours of IDF, we could hardly expect anything nearly as exciting.  While that thought may or may not be true, depending on your perspective, there were definitely a number of interesting announcements and demonstrations throughout the day.  Two of the biggies involved future interfaces - Serial ATA and USB 2.0.

Serial ATA

The vast majority of users out there use the good old ATA, or IDE, interface to hook up their hard drive and CD-ROM. And why not? It’s cheap, readily available, and relatively easy to configure. However, the roots of the standard are over 15 years old and we’re quickly reaching the limits of how far we can extend it with the upcoming ATA100 standard.

This old standard is now being referred to as Parallel ATA since it transfers data in a parallel fashion in other words, multiple bits of data are transferred simultaneously. The shortcomings of this setup are numerous. Simply by being a parallel interface, the pin count is quite high 26 signal pins per channel, plus power and ground. Along the same lines, those 80 conductor ATA66 cables are relatively expensive (from an OEM perspective) and severely clutter a case, which in turn impedes airflow. Further, all ATA designs require 5V tolerant transceivers that complicate things at a time when we’re pushing towards lower and lower voltages throughout our systems. But perhaps the biggest problem is the fact that there is no easy way to push the standard much further it looks like 100-133 MB/s would be the limit unless new signaling were introduced, which of course raises a number of other issues that we won’t get into right now.

Click to enlarge

Click to enlarge

As the name suggests, Serial ATA transfers data serially, which means one bit at a time. The benefit is that pin count of the interface can be significantly reduced, which in turn means cables can be smaller and more efficient. In fact, there are only 4 signal pins and just a few more for power and ground. Since this is a new standard, the opportunity is there to define new signaling with more flexible voltage tolerances.

The Serial ATA specification is being designed with the future in mind, but backwards compatibility is being maintained in a number of ways. Most important is the fact that Serial ATA is software/register compatible with Parallel ATA, which means there is no need to rewrite anything at the driver or OS level.

On the hardware side, conversion devices will initially provide backward compatibility. The devices will convert a Parallel ATA signal from the motherboard to a Serial ATA signal usable by a new hard drive, or vice versa as needed.

The future of Serial ATA is being insured with careful planning. In fact, Serial ATA is expected to last us at least 10 years. Future data rates may be introduced simply by increasing the clock. In fact the roadmap already calls for two follow-ups to Serial ATA, which will double and then double again the transfer rates. Look for Serial ATA 1X to be available in mid 2001 and offer 1.5 gigabits/sec, which translates to an effective 150MB/s. Since Serial ATA will be a point to point protocol, the concept of master/slave no longer exists and each drive will not need to share bandwidth.

Further, unlike the original Parallel ATA spec, Serial ATA is being designed with power management and mobility in mind. With fewer signals in the interface, power consumption is lower. Further, the interface and drive supports native power management that is transparent to the rest of the system. Finally, the connector for 3.5” and 2.5” drives will be identical, which will greatly increase flexibility of mobile and desktop designs.

The group working on Serial ATA features a number of big names, such as Intel, IBM, Maxtor, Seagate, Western Digital, Quantum, Dell, and more. They’ve made it clear that they’re focused on providing a Parallel ATA replacement without adding too many bells and whistles to weigh down the specification and increase the cost of implementation. So don’t expect to use Serial ATA for scanners, printers, etc. that’s what USB and other external interfaces are for.

With that said, they are designing in at least a few advanced features. At least one that is likely to make the spec, although it may be optional, is hot plug support. The opportunity is even there to power drives through the interface cable, which further simplifies configuration. These two features may not be available at launch, but look for it to happen eventually.

Yesterday at IDF, we got to see a demo of Serial ATA in action. Since chipsets and hard drives that support the interface are not yet available, they used the above mentioned converters from a Parallel ATA interface on the motherboard, to a Serial ATA cable, and finally converted back to a Parallel ATA interface on the drive. The demo was simply and consisted of a script that ran a number of Microsoft Office apps.

USB 2.0

USB 2.0 is a simple, but very useful, evolution of the USB 1.1 that we all know and love. The 0.9 spec was finalized in December 1999 and the final 1.0 spec is expected in April of this year. The original goal of USB 2.0 was to have 10 - 20 times the transfer rates of USB 1.1, which offers 12mbps. Well, they did even better than that and ended up with a 40x increase for a total of 480mbps.

The spec was wisely designed to be backwards and forwards compatible, meaning that a USB 2.0 device will work in a USB 1.1 port and vice versa. Of course, a USB 1.1 device won’t speed up any in a USB 2.0 port. Along the same lines, a USB 2.0 device will be limited to 12mbps when plugged into a USB 1.1 port. All cabling and ports are exactly the same.

At IDF, Intel demonstrated a prototype USB 2.0 scanner and an older USB 1.1 printer hooked up to the same system. They scanned in a press release and then printed it on the printer without any problems.

Look for USB 2.0 in the second half of 2000, most likely first in the upcoming i815 Solano chipset.