Hi Mark, As you say, your customers will generate 33% more traffic than those of an ISP which counts national traffic in its data plan. Does this not then mean that your network would be, on average, under 33% more load than a competitor's network? Add to this the fact that such pricing is likely to attract heavy data users (while Xtra's 200mb plans attract grandmothers), and you can quickly see that 24kbits goes a lot further for Xtra (or most other ISPs) than it does for yourself. I've always thought what you were doing with national flatrate pricing was worthy of the New Zealand Order of Merit, given the massive cost differential in NZ between national and international data, but considering Telecom sets the PVC size based on number of customers, rather than the throughput generated by those customers (which in itself strikes me as completely irrational), offering these kind of deals is going to lead to congestion. 24kbits on a per-customer basis is crazy when some customers do 200mb per month and others do 10GB+, It sucks, it really does, but I can't see a solution other than trying to get Telecom to change their PVC-sizing policy, and we all know what the odds of that are. Best of luck, Erin Salmon -----Original Message----- From: Mark Frater [mailto:mfrater(a)quicksilver.co.nz] Sent: 27 April 2006 9:09 p.m. To: 'Simon Allard'; nznog(a)list.waikato.ac.nz Subject: Re: [nznog] QSI / "industry" DSL problems caused by3.5mbit upgrades? Quicksilver's domestic traffic makes up about a 3rd of its total traffic, which is probably not that different from other ISPs. It doesn't appear to change the average persons behaviour. People don't tend to visit trademe any more or any less just because it doesn't count towards their datacap, and most of the peer to peer traffic comes from offshore.

From an average customer perspective, it just means that their effective datacap is 25% higher with Quicksilver than other ISPs. e.g. a QSI customer on a 10GB international plan will on average have downloaded 13.3GB (domestic and international) before their international speed is throttled.

Mark -----Original Message----- From: Simon Allard [mailto:simon.allard(a)staff.ihug.co.nz] Sent: Thursday, 27 April 2006 4:39 p.m. To: nznog(a)list.waikato.ac.nz Subject: Re: [nznog] QSI / "industry" DSL problems caused by3.5mbit upgrades?

With this service limitation I can easily see how Quicksilver could potentially be affected more than Ihug, Iconz, Xtra and Orcon. I would guess that the average user profile on Quicksilver would tend towards higher bandwidth usage whereas the typical Xtra (etc) user would probably use less bandwidth.

Unfortunately if this view I have is correct then I don't see an end to problems unless either the average Quicksilver user profile changes -or- Telecom changes the PVC limitation.

I am curious to know what effect offering free national traffic has on the QSI DSL Network. In my experience offering free anything in this environment Telecom has created causes nothing by problems. _______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog -- No virus found in this incoming message. Checked by AVG Free Edition. Version: 7.1.385 / Virus Database: 268.5.0/325 - Release Date: 26/04/2006 -- No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.1.385 / Virus Database: 268.5.0/325 - Release Date: 26/04/2006 _______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog

Hi Erin, Fortunately what you have assumed is not the case in reality. You could hypothesise that if all our customers reached their datacap each month, and all customers of another ISP reached their caps each month, and the caps were the same, and we both had the same number of customers and no customers on either ISP bought additional datablocks, then you could say that the monthly average utilisation of our network would be 25% higher than the competitors network. In reality, only a small number of our customers reach their datacaps, and the average number of gigabytes that our customers download is well below many of our competitors (such as those offering unlimited download plans!) Offering what effectively amounts to a 13G plan instead of a 10G plan makes little difference makes little difference to the monthly traffic profiles. You could argue that near the end of the month, when some of a competitors customers are starting to get capped and ours are not, that the competitor would see their network utilisation reduce slightly. But more often than not, people who are capped just go buy another datablock and continue to hammer their download. I take your point about Xtra though.. They must have most of the "Ma and Pa" users, who do very little other than generate revenue for Telecom. Mark -----Original Message----- From: Erin Salmon - Unleash [mailto:erin(a)unleash.co.nz] Sent: Thursday, 27 April 2006 10:13 p.m. To: nznog(a)list.waikato.ac.nz Subject: Re: [nznog] QSI / "industry" DSL problems caused by3.5mbit upgrades? Hi Mark, As you say, your customers will generate 33% more traffic than those of an ISP which counts national traffic in its data plan. Does this not then mean that your network would be, on average, under 33% more load than a competitor's network? Add to this the fact that such pricing is likely to attract heavy data users (while Xtra's 200mb plans attract grandmothers), and you can quickly see that 24kbits goes a lot further for Xtra (or most other ISPs) than it does for yourself. I've always thought what you were doing with national flatrate pricing was worthy of the New Zealand Order of Merit, given the massive cost differential in NZ between national and international data, but considering Telecom sets the PVC size based on number of customers, rather than the throughput generated by those customers (which in itself strikes me as completely irrational), offering these kind of deals is going to lead to congestion. 24kbits on a per-customer basis is crazy when some customers do 200mb per month and others do 10GB+, It sucks, it really does, but I can't see a solution other than trying to get Telecom to change their PVC-sizing policy, and we all know what the odds of that are. Best of luck, Erin Salmon -----Original Message----- From: Mark Frater [mailto:mfrater(a)quicksilver.co.nz] Sent: 27 April 2006 9:09 p.m. To: 'Simon Allard'; nznog(a)list.waikato.ac.nz Subject: Re: [nznog] QSI / "industry" DSL problems caused by3.5mbit upgrades? Quicksilver's domestic traffic makes up about a 3rd of its total traffic, which is probably not that different from other ISPs. It doesn't appear to change the average persons behaviour. People don't tend to visit trademe any more or any less just because it doesn't count towards their datacap, and most of the peer to peer traffic comes from offshore.

From an average customer perspective, it just means that their effective datacap is 25% higher with Quicksilver than other ISPs. e.g. a QSI customer on a 10GB international plan will on average have downloaded 13.3GB (domestic and international) before their international speed is throttled.

Mark -----Original Message----- From: Simon Allard [mailto:simon.allard(a)staff.ihug.co.nz] Sent: Thursday, 27 April 2006 4:39 p.m. To: nznog(a)list.waikato.ac.nz Subject: Re: [nznog] QSI / "industry" DSL problems caused by3.5mbit upgrades?

With this service limitation I can easily see how Quicksilver could potentially be affected more than Ihug, Iconz, Xtra and Orcon. I would guess that the average user profile on Quicksilver would tend towards higher bandwidth usage whereas the typical Xtra (etc) user would probably use less bandwidth.

Unfortunately if this view I have is correct then I don't see an end to problems unless either the average Quicksilver user profile changes -or- Telecom changes the PVC limitation.

I am curious to know what effect offering free national traffic has on the QSI DSL Network. In my experience offering free anything in this environment Telecom has created causes nothing by problems. _______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog -- No virus found in this incoming message. Checked by AVG Free Edition. Version: 7.1.385 / Virus Database: 268.5.0/325 - Release Date: 26/04/2006 -- No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.1.385 / Virus Database: 268.5.0/325 - Release Date: 26/04/2006 _______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog _______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog -- No virus found in this incoming message. Checked by AVG Free Edition. Version: 7.1.385 / Virus Database: 268.5.0/325 - Release Date: 26/04/2006 -- No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.1.385 / Virus Database: 268.5.0/325 - Release Date: 26/04/2006

Hey Joe, I think the simple answer is that its not just Quicksilver, as postings to numerous blogs and mailing lists would suggest. Here is the lastest from ISPANZ - http://www.ispanz.org.nz/press_release14 However, I think that Quicksilver may have been subjected to a bit of a double-whammy (technical term) this week given that we are at the end of out last Telecom UBS PVC upgrade provisioning cycle, and our PVC will be upgraded in the next couple of days "to provide for forecasted UBS customer growth over the next 3 months" (to quote the official line to Telecom). As some have suggested, we could say to Telecom that we forecast an additional 5000 customers in the next month, but it’s a pretty sad state of affairs if you have to lie to your major supplier in order to compete with them on an equal footing. (ok I guess we all already know the broadband IS a sad state of affairs in NZ.) In my mind, if we're paying for 155Mbit/s ATM circuit, then we should be getting 155Mbit/s (less cell tax) and not be subject to this extra throttle point of 24kbit/s per customer, when any contention ratios are theoretically taken care of further back into the Telecom network. And Xtra is not subject to this extra throttle, so why should we be? By the way, the figure of 24kbit/s apparently includes IP headers, so the actual throughput is even worse. Mark -----Original Message----- From: Joe Abley [mailto:jabley(a)hopcount.ca] Sent: Friday, 28 April 2006 3:46 a.m. To: Mark Frater Cc: 'Erin Salmon - Unleash'; nznog(a)list.waikato.ac.nz Subject: Re: [nznog] QSI / "industry" DSL problems caused by3.5mbit upgrades? Hi Mark! On 27-Apr-2006, at 05:50 , Mark Frater wrote:

In reality, only a small number of our customers reach their datacaps, and the average number of gigabytes that our customers download is well below many of our competitors (such as those offering unlimited download plans!)

Matt's ad-hoc survey seemed to suggest that it's just qsi customers suffering from edge congestion and peak times, and not those of other ISPs. If qsi's trunks to Telecom are provisioned identically to those at other ISPs, and if you are correct in that the average customer use profile is near-identical, what is left that might explain the observed difference in performance? Joe -- No virus found in this incoming message. Checked by AVG Free Edition. Version: 7.1.385 / Virus Database: 268.5.0/325 - Release Date: 26/04/2006 -- No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.1.385 / Virus Database: 268.5.1/326 - Release Date: 27/04/2006