W3C home > Mailing lists > Public > public-device-apis@w3.org > May 2013

RE: [network information] adding latency; defining how to obtain latency/bandwidth estimates or providing precision - comment?

From: Mandyam, Giridhar <mandyam@quicinc.com>
Date: Wed, 8 May 2013 13:57:53 +0000
To: "Frederick.Hirsch@nokia.com" <Frederick.Hirsch@nokia.com>, "Diana.Cheng@vodafone.com" <Diana.Cheng@vodafone.com>
CC: "public-device-apis@w3.org" <public-device-apis@w3.org>
Message-ID: <CAC8DBE4E9704C41BCB290C2F3CC921A164624AD@nasanexd01h.na.qualcomm.com>
I tend to +1 on Diana's view.

Moreover, developers can make their own measurements using PerformanceTiming (http://www.w3.org/TR/navigation-timing/).  For latency to be included as part of Network Info, there should be demonstrable value over features that have already been standardized.


-----Original Message-----
From: Frederick.Hirsch@nokia.com [mailto:Frederick.Hirsch@nokia.com] 
Sent: Wednesday, May 08, 2013 6:31 AM
To: Diana.Cheng@vodafone.com
Cc: Frederick.Hirsch@nokia.com; public-device-apis@w3.org
Subject: Re: [network information] adding latency; defining how to obtain latency/bandwidth estimates or providing precision - comment?


Your points are well taken.

Can proponents please share use cases on the list? I'm assuming latency is of interest due to user interface responsiveness but more detail would be welcome. 

Clearly implementation details matter, raising the question of accuracy and timeliness of the information versus battery drain.

regards, Frederick

Frederick Hirsch

On May 8, 2013, at 9:13 AM, ext Cheng, Diana, Vodafone Group wrote:

> What is the use case for this? Many developers don't even realise the 
> difference between latency and bandwidth...
> Also, latency on mobile can vary widely: when the radio has been IDLE 
> and it needs to negotiate the link, you have an up to 2000ms added. If 
> measured then it isn't an indication of later RTTs. Mobile latency is 
> also shaped by carriers, and can change within a couple of seconds, in 
> order to, for example, give momentary priority to certain types of 
> traffic - like latency-critical M2M applications, or others.
> Moreover, constantly measuring these parameters keeps the radio in 
> full power, and that is ©well, bad.
> Thanks,
> Diana.
> On 07/05/2013 20:45, "Frederick.Hirsch@nokia.com"
> <Frederick.Hirsch@nokia.com> wrote:
>> [sent not as chair]
>> On our 24 April DAP teleconference we discussed the Network 
>> Information API [1] noting that both latency and bandwidth are useful 
>> information, and that in many cases latency information may even be 
>> more interesting than bandwidth information.
>> We also noted that it would be useful if different implementations 
>> shared a common means of providing the estimates, so that 
>> applications could have a basis for interpretation despite platform.
>> That said, I expect such estimates may be implementation dependent - 
>> in that case it may be useful to have an indication of precision.
>> Thus we have two questions:
>> 1. Is there general agreement that it would be useful to add latency 
>> to the Connection Web IDL of the Network Information API
>> 2. Would it be useful and possible to add information to the 
>> specification on how to provide latency and bandwidth estimates? If 
>> not, should we be including some sort of precision for the each of 
>> the estimates?
>> Thanks
>> regards, Frederick
>> Frederick Hirsch
>> Nokia
>> [1] http://www.w3.org/TR/2012/WD-netinfo-api-20121129/
>> for Tracker, this should complete ACTION-629
Received on Wednesday, 8 May 2013 13:58:23 UTC

This archive was generated by hypermail 2.3.1 : Monday, 23 October 2017 14:53:59 UTC