sata io SATA IO announces DevSleep for reduced power draw of SATA devices

We seem to be living in the age of power frugality when it comes to mobile computing devices and although this is by no means a bad thing in most cases, we're starting to wonder if the technology companies aren't looking in the wrong direction sometimes. Now the SATA-IO is the latest organization to join with a new power saving feature called DevSleep which is meant to help reduce the power of SATA devices.

We seem to be living in the age of power frugality when it comes to mobile computing devices and although this is by no means a bad thing in most cases, we're starting to wonder if the technology companies aren't looking in the wrong direction sometimes. Now the SATA-IO is the latest organization to join with a new power saving feature called DevSleep which is meant to help reduce the power of SATA devices.

According to Microsoft and Intel the two most power hungry devices in a notebook is the hard drive and the display and it's clear that the SATA-IO is trying to address the first one of these two categories. What DevSleep enables is the ability for a SATA device such as an SSD or hard drive to more or less power itself off entirely when not in use. This is quite different from the various slumber modes currently available for SATA devices.

One problem with DevSleep is that new device and host controllers would need to be made that supports DevSleep, as the PHY and other parts of the storage device would be completely powered off when the device is in DevSleep mode. In DevSleep mode a SATA device is meant to draw less than 5mW, yet have a response time of less than 20ms. This is a lot quicker than resuming a notebook from standby and should equate to a similar power draw. A SATA device today would draw in the region of 50-100mW in its lowest slumber mode, yet still take 10ms to respond, so the trade-off here is very small in terms of what the user would notice in terms of delay, yet big in terms of power saved.

An added bonus is that the host controller PHY would also be powered down, saving further power. The problem here is that it wouldn't work with today's SATA controller, as a DevSleep side-band signal would be needed to allow the host and device to communicate as to when to go to sleep and when to wake up. As such it's not likely that we'll see DevSleep support in any SATA devices this year, although the information provided by the SATA-IO doesn't really explain what is needed to add the extra side-band signalling, so it's entirely possible that this is a very minor addition that could easily be implemented in current SATA hardware.

Source: SATA-IO