- AT&T Application Resource Optimizer (ARO)
- Additional device profiles & 4 state RRC model?
07.29.2012 09:01 PM - edited 07.29.2012 09:04 PM
Developers want to test and target for the broadest market. So while it certainly makes sense that you have focused on devices that AT&T supports, it would be great if there was a larger library of device profiles available to load. Are you aware of such a library or have any plans to populate yours with more profiles?
My second question is similar: you provide network models for 3G & LTE. However, your 3G model doesn't include CELL_PCH (presumably because this isn't implemented on the AT&T network). Do you have any plans to extend the RRC models to include 4 state WCDMA models? Alternatively, do you know anyone who has extended your existing model?
07.30.2012 06:28 AM
You ask severl great questions in this post. I'll do my best to go through them.
1. Additional devices.
In the tool, you may create additional profiles for devices. The ONLY difference behind the various 3G devices are the power levels. Because the HTC Aria has a different radio power than the Samsung Infuse (i997), you will see a difference in power output when you apply the 2 models. The only thing changing is the power math. Our main goal with ARO is not to measure power levels (yes we measured these manually) of 100s of devices, but is to give you as a developer an idea of how much power is being drained out.
2. CELL_PCH. PCH runs at about 1% of DCH. With DCH = 800-1000mW, that puts CELL_+PCH at 8-10mW. You can essentially model that as IDLE (power =0). We do have the addition CELL_PCH in our deveklopment backlog however (as wella s Fast Dormancy).
While the ARO model is not 100% accurate to the WCDMA 4 state model, what is important to realize is that these timers are fluid, and change as needed by our network engineers. Our main goal with ARO is to proivde developers with INSIGHT into the state machine. In speaking to developers for over 18 months, only about 5-10% have ever heard of the state machine. In many cases, justr knowing how the state machine affects the battery is enough to realize that managing network connections in a smart way is really important.
- of 1