X-List-Received-Date: Fri, 25 Jul 2014 15:46:37 -0000 --Apple-Mail=_85332351-916D-4F5B-92B5-45687EE8479E Content-Type: multipart/alternative; boundary="Apple-Mail=_9E651F6D-9FCC-40C5-B3F9-2CAB1AB21FE4" --Apple-Mail=_9E651F6D-9FCC-40C5-B3F9-2CAB1AB21FE4 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=windows-1252 Hi Neil, > I have a deep worry over this style of single point measurement - and = hence speed - as an appropriate measure. We know, and have evidence, = that throughput/utilisation is not a good proxy for the network = delivering suitable quality of experience. We work with organisation = (Telco=92s, large system integrators etc) where we spend a lot of time = having to =93undo=94 the consequences of =93maximising speed=94. Just = like there is more to life than work, there is more to QoE than speed. I completely agree with this: those of us who have spent the time to = ponder the physics of the problem have come to understand it in its full = glory (and complexity). We know that a single number ain't the answer. =20 But one of my other goals is to increase the awareness of the = bufferbloat problem. People *do* use these speed test services, despite = their inaccuracy. If those sites could include some form of measurement = about the latency (and its change) under load, it makes it easier to = describe the problem to others. The long-term solution is, of course, to get router vendors to realize = there's a problem and then respond to market pressures to fix it. I = gained a lot of insight from http://apenwarr.ca/log/?m=3D201407#11 - it = has a description of the difficult economic justification for selling a = "good router". But where that author was trying to start a company, = we're in a different position.=20 Good latency info in a popular speed test website elevates its = importance to the general public. It moves you out of the "nutcase" = category ("What's this bufferbloat stuff this guy's talking about?"), = into a concerned customer who's offering a useful observation. And it = gives credibility when you bug providers. Examples: - A lot of people reflexively check speedtest.net when they check into a = hotel, and then post/tweet the results. If the results include min/max = latency, then they can begin to comment to the hotel (where they may = even have been charged for the service) when things aren't good.=20 - I frequently ride a bus to Boston that offers free wifi. I already = know they're badly bloated (>20 seconds(!)). With a tool like this, it's = easier to begin a conversation with the operations people that lets them = put pressure on their own router vendor.=20 - Home and commercial users can use these values to tell their ISPs that = there's a problem, and the expense of fielding those support calls = provides some incentive to address the problem. - Heaven forbid one vendor/provider snatch up the idea and tout it as a = competitive advantage... :-) We've shown that there is a straightforward fix to the problem. With = increased awareness from their customers, I hold out hope that we can = begin to change the world. Best, Rich PS Thanks to Sebastian for the comments, and Martin for those links. --Apple-Mail=_9E651F6D-9FCC-40C5-B3F9-2CAB1AB21FE4 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=windows-1252 Hi = Neil,

I have a deep worry over this style of = single point measurement - and hence speed - as an appropriate measure. = We know, and have evidence, that throughput/utilisation is not a good = proxy for the network delivering suitable quality of experience. We work = with organisation (Telco=92s, large system integrators etc) where we = spend a lot of time having to =93undo=94 the consequences of =93maximising= speed=94. Just like there is more to life than work, there is more to = QoE than speed.

I completely = agree with this: those of us who have spent the time to ponder the = physics of the problem have come to understand it in its full glory (and = complexity). We know that a single number ain't the = answer.

But one of my other goals is to = increase the awareness of the bufferbloat problem. People *do* use these = speed test services, despite their inaccuracy. If those sites could = include some form of measurement about the latency (and its change) = under load, it makes it easier to describe the problem to = others.

The long-term solution is, of course, = to get router vendors to realize there's a problem and then respond to = market pressures to fix it. I gained a lot of insight from http://apenwarr.ca/log/?m=3D= 201407#11 - it has a description of the difficult economic = justification for selling a "good router". But where that author was = trying to start a company, we're in a different = position.

Good latency info in a popular = speed test website elevates its importance to the general public. It = moves you out of the "nutcase" category ("What's this bufferbloat stuff = this guy's talking about?"), into a concerned customer who's offering a = useful observation. And it gives credibility when you bug providers. = Examples:

- A lot of people reflexively check = speedtest.net when they check into = a hotel, and then post/tweet the results. If the results include min/max = latency, then they can begin to comment to the hotel (where they may = even have been charged for the service) when things aren't = good.

- I frequently ride a bus to Boston that offers = free wifi. I already know they're badly bloated (>20 seconds(!)). = With a tool like this, it's easier to begin a conversation with the = operations people that lets them put pressure on their own router = vendor.

- Home and commercial users can use these values = to tell their ISPs that there's a problem, and the expense of fielding = those support calls provides some incentive to address the = problem.

- Heaven forbid one vendor/provider snatch up the = idea and tout it as a competitive advantage... = :-)

We've shown that there is a straightforward = fix to the problem. With increased awareness from their customers, I = hold out hope that we can begin to change the = world.

</utopian = rant>

Best,

Rich

PS Thanks to Sebastian for the comments, and Martin = for those links.

= --Apple-Mail=_9E651F6D-9FCC-40C5-B3F9-2CAB1AB21FE4-- --Apple-Mail=_85332351-916D-4F5B-92B5-45687EE8479E Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename=signature.asc Content-Type: application/pgp-signature; name=signature.asc Content-Description: Message signed with OpenPGP using GPGMail -----BEGIN PGP SIGNATURE----- Comment: GPGTools - https://gpgtools.org iQEcBAEBCgAGBQJT0nvYAAoJEH4agC/0z73/Gg0H/0tGysbszcbSOLBe/NrZi83g 4kF0gcoQMrI/zkgL8TSnSXPGyBQB7Y8qHhs87o19kJyA/rCOLtzoFU1Y7IDVeYkh /Xx+AYPoucTxOhdKcq3hnQR3TqKF8pBuDPIdQvK0Dm399UdZxanlCyMivjzjGMMT lZkriRn1iEcdlGD+AKMBmqV2tUI2N/QquukJoyPV2BjfmRmUCYjZLWHHCRopo6Oy qV8NdWWrlxwMIDN9YPLAj9h+FRsDkK+1GAwmiBI5AURFEBxxGaNW1yPTgqJcclmq Yoy7Z3IqRZ5beJCVkSoiim1FHXdumYYe8boj8AHWOOQCNYs4vBBBGolMyh5K+2o= =HZdP -----END PGP SIGNATURE----- --Apple-Mail=_85332351-916D-4F5B-92B5-45687EE8479E--