Closed Bug 993975 Opened 12 years ago Closed 8 years ago

[Sora][Wi-Fi]Establishing the data connection above 3 seconds

Categories

(Firefox OS Graveyard :: Wifi, defect, P2)

defect

Tracking

(tracking-b2g:backlog)

RESOLVED WONTFIX
tracking-b2g backlog

People

(Reporter: sync-1, Unassigned)

Details

(Keywords: perf, Whiteboard: [c=hardware p= s= u=])

Mozilla build ID:20140323004002 DEFECT DESCRIPTION: Establishing the data connection above 3 seconds. REPRODUCING PROCEDURES: 1,Connect the AP successfullly before->Open the Wi-Fi ->count the establishing the data connection time.-> The time is between 6,7 seconds -> KO EXPECTED BEHAVIOUR: The Wi-Fi scan SHALL not introduce more than 3 seconds delay to establishing the data connection. ASSOCIATE SPECIFICATION: TEST PLAN REFERENCE: TOOLS AND PLATFORMS USED: USER IMPACT: REPRODUCING RATE: For FT PR, Please list reference mobile's behavior:
Hi - The time needed to connect to the AP also depends on the AP and the whole infrastructure, as you can see from the following video, in here it only takes me less than 4 seconds to connect to the AP: https://www.youtube.com/watch?v=xUzR_cnlk80&feature=youtu.be So suggest to test on more AP to calculate the average connection time.
Flags: needinfo?(sync-1)
Our tester is going to working this average time, I'll feedback test result later.
Some steps in gonk takes much time, like load driver, start wpa_supplicant, load wifi network settings, and can only be optimized by partners. Take unagi for example, these steps takes at least 3 seconds. To interact with gecko, the time will be longer. Also, getting IP from DHCP after wifi connected takes 200ms ~ 10seconds, which we can not control. Bug 994564 try to parallelize some of these step, might make the time shorter. But I think 3 seconds is still not likely a possible target.
Keywords: perf
Flags: needinfo?(sync-1)
blocking-b2g: --- → backlog
Ravi, Per triage, please confirm competitive requirements for this operation and update this bug saying whether our performance needs to be improved or if we are currently meeting our goal. See comment 3 for details on current limitations. Thanks, Mike
Flags: needinfo?(rdandu)
Whiteboard: [c=hardware p= s= u=]
Can we find out how much Android device takes on the same device HW, and AP.
Flags: needinfo?(jsmith)
Flags: needinfo?(rdandu)
(In reply to rdandu from comment #5) > Can we find out how much Android device takes on the same device HW, and AP. Tony - Do we have comparable hardware to test this on to compare the behavior on Android vs. FxOS?
Flags: needinfo?(jsmith) → needinfo?(tchung)
(In reply to Jason Smith [:jsmith] from comment #6) > (In reply to rdandu from comment #5) > > Can we find out how much Android device takes on the same device HW, and AP. > > Tony - Do we have comparable hardware to test this on to compare the > behavior on Android vs. FxOS? I dont think thats the right approach to take. Instead, i would put it on Reporter to compare against the Alcatel One Touch Fire running 1.3, and see what it compares there. This way they're also testing against the same network as the Sora, and can compare if there's a core problem with the gecko code or hardware itself.
Flags: needinfo?(tchung)
blocking-b2g: backlog → ---
link all Fire C (codename: Sora) bugs to a meta one.
Firefox OS is not being worked on
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.