[Advisors] It’s no surprise, but it’s official: Europeans don’t get the broadband speeds they pay for

[Gary W Kenward]

Pandora's box: the difficulty lies in defining "broadband speeds". 

As benchmarks, the algorithms developed by SamKnows look fine. However, I did notice one feature of the SamKnows benchmark: ISPs typically advertise the bit transmission rate over the channel and SamKnows tests measure user payload throughput (over TCP and FTP).  User data transfer speeds will always be less then the bit transmission rate due to protocol overheads. The overhead for TCP/IP and FTP/IP is significant. The impact of this overhead on the user experience is highly variable, depending on a number of technical factors.

Industry benchmarks for computers and digital communications have always been controversial because of the many variables involved and the wide range of interpretations. 


Wired downlink "speeds" are largely going to be effected by protocol overheads, although there is always the possibility of a congestion occurring an under-provisioned router or link, or, a problem I often encounter with my cable provider, soft failure of the physical link. Wireless downlink speeds are somewhat similar, although they suffer from the variability of the RF channel.

Uplink speeds are another matter: DSL and FTTx provide dedicated bandwidth to the end user; cable and fixed point wireless provide multi-access channel shared along the "last mile" (e.g. with your neighbours)


Benchmarking cellular data is a completely different conundrum, as the wireless channel introduces many more variables. The industry standard is to advertise the peak theoretical bit rate for a sector (and sometimes for the cell). Even assuming maximum RF signal, unless you are the only person on the channel a user is highly unlikely to experience a sustained rate even near to the peak theoretical. And then there are the overheads....


I believe a nation wide set of government sanctioned benchmarking standards and binding regulations (i.e. CRTC) would be beneficial to all Canadians. There is prior history regarding benchmarks that suggests this will be a difficult to achieve.

The alternative is a grass roots effort: the SamKnows web site asks for volunteers to host benchmarking devices in their homes and businesses. This initiative does not appear to be active yet (the firmware is not yet available to the public, but you can sign up to a list).

Regards,
Gary


BTW: To give you an idea of where I am coming from: I have worked on Internet throughput and general data transport performance since the '90s. Part of my responsibilities as Data Application Architect for Nortel was to characterize access channel capacity and performance. I never agreed to the marketing spin (let's call it that) put on advertised numbers, which didn't make me very popular. To be fair to Nortel, and the cellular industry as a whole, citing peak instantaneous bit transmission rate at the transmitter is basically the default industry standard.


On 2013.06.27, at 1:05 PM, Evan Leibovitch <evan at telly.org> wrote:

> I don't know about stats, but I've been on both Rogers and Bell. In my experience, I'm overpaying but they give what they promise.
> 
> I regularly use http://206.47.199.107/ (provided by Bell) and speedtest.net.
> 
> 
> 
> 
> On 27 June 2013 00:52, Garth Graham <garth.graham at telus.net> wrote:
> Does anyone have comparable stats for Canada?
> 
> It’s no surprise, but it’s official: Europeans don’t get the broadband speeds they pay for
> http://thenextweb.com/eu/2013/06/26/its-no-suprise-but-its-official-europeans-dont-get-the-broadband-speeds-they-pay-for/
> 
> GG
> 
> _______________________________________________
> Advisors mailing list
> Advisors at tc.ca
> http://victoria.tc.ca/mailman/listinfo/advisors
> 
> 
> 
> 
> -- 
> Evan Leibovitch
> Toronto Canada
> Em: evan at telly dot org
> Sk: evanleibovitch
> Tw: el56
> _______________________________________________
> Advisors mailing list
> Advisors at tc.ca
> http://victoria.tc.ca/mailman/listinfo/advisors

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://victoria.tc.ca/pipermail/advisors/attachments/20130627/002fa19b/attachment.html>