Drop list consultation

[BMW7series]

I hope you're not implying that people like myself who offer dropcatching systems are 'operating multiple registrars'. There is a big difference between multiple people using the same script (where many tagholders are using the same scripts they bought or lease) and somebody with multiple tags under one control which is against nominet rules.

This kind of misunderstanding has already led to one incorrect veiled accusation on this forum. I don't 'operate' any registrar except my own.

Hello and sorry if i riled you because that wasn't the intention.

"Those who are operating multiple registrars for drop catching purposes now [cut] by offering DAC hosting services to others..."

I'll clarify that I meant facilitating the operation of multiple registrars... I hope that is clearer. I haven't examined every DAC hosting service in great detail to become familiar with how each one operates. They may all differ with their modes of operation.

I do think the likes of those will have some interesting decisions to make if Nominet do decide to publish exact and unique drop times for each domain name and have explained why I think that above. At the moment a great many domain name availability checks need to be sent to determine the availability of any specific domain name due to be deleted by the registry prior to an EPP Create. If all that is made redundant, meaning success comes down to efficient code and perfect timing, some registrars who solely exist to engage in drop catching may find themselves without a host if that host can now do very well for themself in-house.

Why would the best developers give away or lease or sell their new efficient and time perfected scripts to anyone else when they can register all or most of the domain names they wanted themselves?

However until nominet tell us what they have decided after this charade of a consultation we can't really say how everyone will adapt. Hopefully I'll be retired, or dead, by then anyway. We do know from past experience with price hikes and uk introductions that they only care about their own, and their top registrars', bottom line. It's easy for them to funnel the money from them - impossible from us. Just out of interest what made you join and post straight in this thread? Are you related to nominet? Not an accusation just wondering if they are sending in someone to butter us up

I don't enjoy posting on forums ordinarily. It commands a lot of time and I do enjoy being involved in many other things elsewhere which compete for my time. This topic interested me but I think I've probably said almost everything I can add.

 

[RobM]

It depends on how random it really is I guess. If it is random then of course people would lease their software - better to have lots of people buying or paying a license and have guaranteed income and *still* the same low chance of success. However nominet showed in the RoR that they lacked the ability and the infrastructure to actually release domains at the time stated so it'll just be a matter of time until people figure out their flaws. All they will be doing is removing the stress from the DAC and throwing extra stress on the EPP. This wouldn't make much sense though - they will need to find a way to give their top registrars the edge - that is what most people are interested in I think. How 'equal' will the whole system be. The automaton was far fairer and less resource intensive than the DAC - they scrapped that. For a while they got to take advantage of the EPP via certain avenues but now everybody is in on their problems. So now they will scrap this. The fairer a system becomes the less they want it.

 

[BMW7series]

It depends on how random it really is I guess.

From what I've read the intention is it wouldn't be random anymore. Firstly a daily drop list would negate the need for list building which is a waste of resources due to huge numbers of DNS queries and domain name availability checks sent.

A daily drop list with exact drop timestamps, or as near to exact as possible, would negate the need to check for any domain name availability repeatedly. Developers would come to know the margin of error +/- the published time and if that varied they could continue to compensate for that variance. That's how they would set themselves apart from others.

If it is random then of course people would lease their software - better to have lots of people buying or paying a license and have guaranteed income and *still* the same low chance of success.

I feel that may be the case when sending repeated domain name availability checks is necessary like it is currently. I'm not sure I think it will remain the case if exact and unique domain name deletion timestamps are published and the margin of error is tiny.

However nominet showed in the RoR that they lacked the ability and the infrastructure to actually release domains at the time stated so it'll just be a matter of time until people figure out their flaws.

I feel that was materially different due to the large number of domain names that were released at exactly the same time each day, for five days, and where a great many were contested for. Far fewer domain names are contested each day compared to during the ROR and the proposal appears to suggest unique drop times rather than everything at once.

All they will be doing is removing the stress from the DAC and throwing extra stress on the EPP.

Can't be sure. Depends how many EPP Create requests are sent at the moment for contested domain names by all registrars who participate and notice a domain names availability. If drop timestamps are published my thinking is that number of EPP creates would reduce.

This wouldn't make much sense though - they will need to find a way to give their top registrars the edge - that is what most people are interested in I think. How 'equal' will the whole system be. The automaton was far fairer and less resource intensive than the DAC - they scrapped that. For a while they got to take advantage of it via certain avenues but now everybody is in on their problems. So now they will scrap this. The fairer a system becomes the less they want it.

I am not of the view that there is a desire or even a requirement to give any registrar more EPP creates than others. As I said previously this is not the ROR with hundreds of desirable domain names being deleted at the same time. How many dropped domain names are generally reregistered during an average day? 100 or fewer? If 100, one registrars sending 5 EPP Create requests for each would still only use 500 creates. Far fewer than the total current limit per registar.

 

[stitchbob]

If the release becomes similar to ROR, the barrier to entry for chasing would become much lower. Right now, you need to perform a DAC check, get a response, and send an EPP request to register the domain, while competing with numerous others doing the same thing. That requires a very fast connection as close as possible to Nominet's servers, and a fast server to actually process the request in time. A known drop time means you simply need a consistent response time and a bit of trial and error to determine the optimal time to make your registration attempt. A $5 Digital Ocean droplet could achieve this easily.

Of course, the actual catch is only a small part of operating a catch service. If you're not getting enough catches because of the amount of competition, the administration would quickly become more trouble than it's worth.

 

[BMW7series]

If the release becomes similar to ROR, the barrier to entry for chasing would become much lower. Right now, you need to perform a DAC check, get a response, and send an EPP request to register the domain, while competing with numerous others doing the same thing. That requires a very fast connection as close as possible to Nominet's servers, and a fast server to actually process the request in time. A known drop time means you simply need a consistent response time and a bit of trial and error to determine the optimal time to make your registration attempt. A $5 Digital Ocean droplet could achieve this easily.

Of course, the actual catch is only a small part of operating a catch service. If you're not getting enough catches because of the amount of competition, the administration would quickly become more trouble than it's worth.

I agree with you and disagree with you.

I am not of the opinion, from reading what has been published by Nominet so far, that a one time drop each day of everything is the proposal. It might be an option but the suggestion of publishing drop times seems to indicate that each domain name scheduled to drop would be allocated a unique timestamp for when that would occur. This apparently already occurs now with MMX gTLD domain names that Nominet is the RSP for.

I agree with you that there should be no more need for optimally located servers and fast Internet connections with the fewest hops to the DAC because registrars engaged in drop catching would no longer be waiting for DAC responses. Based on the idea of published domain name drop times, one would discover ones routing time to Nominet and could factor that in together with the published drop time and with any offset. The EPP create could be sent from a server in another continent.

Another factor might be a requirement for the developer to be more hands on during the day, monitoring timing offset prior to any attempts to register desired domain names at their published drop times, to see if anything was varying.There might be no more loading lists just before midnight and returning again late in the evening to do it again, leaving the system to do its thing in between. This extra hands on requirement is an added cost.

If the best developers end up being very successful with just a single registrar quota of EPP Creates I am not sure why they would want to give others the opportunity to use their code. Hypothetically speaking, if you could catch ninety percent of everything you attempted to register *today* with a new piece of code would you really let others have a share in that or would you make hay while the sun shone?

 

[RobM]

Expect to see hundreds of tags 'appear' if that's the case although there will come a point where the cost for the % chance gained will be outweighed by the relatively low value (and declining) of co.uk. Personally I think nominet are completely missing the point and should be doing what they were supposed to do all along - promote and maintain the uk namespace. It appears to me they've just given up on it - I guess there's more money to be made doing other things. After all they can always hitch a ride on the gravy train to one of the companies they are funding already....

 

[stitchbob]

Hypothetically speaking, if you could catch ninety percent of everything you attempted to register *today* with a new piece of code would you really let others have a share in that or would you make hay while the sun shone?

Even if you had the absolute fastest script and could guarantee catching every domain you were chasing, you're currently still at the whim of the drop order lottery. It's very common to miss everything because you chose the last to drop as your first to chase.

Published drop times would remove the drop order lottery, but I don't believe that would be better for chasers. If we assume that there are 50 chasers all chasing 10 domains on equal footing, the drop order means you're competing with an average of 5 others for the first domain, with the number of competitors increasing as the list shortens, with all 50 of you finally chasing the last to drop. With the drop order published, all 50 of you are chasing the same domains at the same time. Instead of having a 20% chance lengthening to 2%, every domain is 2%.

 

[BMW7series]

Expect to see hundreds of tags 'appear' if that's the case although there will come a point where the cost for the % chance gained will be outweighed by the relatively low value (and declining) of co.uk.

Will such hypothetical "hundreds of registrars ('tags')" come with the technical ability to develop a system to be able to successfully register very many/any contended domain names at their published drop times? If nobody who is an existing developer of drop catching solutions is prepared to share their code, these registrars would have to source their own developer who would need to be as good as those already reaping success from a future drop cycle that was based on published domain name drop times.

Such a proposed drop cycle isn't based on the use of resources like the current one is but is based on efficient code and great timing. Two things that can't simply be purchased, but require a developer with the skill to produce and maintain the code and how many of them operate currently? I suspect a significant number of registrars that exist to engage in drop catching are using the same codebase. Only a small number are operating something unique, whether or not they lease/share it.

Such a proposal would likely require an entire rewrite of everyones drop catching systems, to take into account syncing to Nominet's time clock and noticing any variances. Those that can do that could do very well with just their single registrar and current daily EPP create quota. DAC hosts might decide to close everything down and inform all their registrar clients that they can bid on the domain names they will be catching on just a single registrar tag at auction like everyone else. No need for any other relationship.

Even if you had the absolute fastest script and could guarantee catching every domain you were chasing, you're currently still at the whim of the drop order lottery. It's very common to miss everything because you chose the last to drop as your first to chase.

Yes I understand this. Forget that aspect and just presume you got the order right all the time and always caught everything. I know it's hypothetical but I believe anyone in that position today wouldn't be sharing their abilities out because there would be no need. They'd be telling anyone to see them at the auctions.

Published drop times would remove the drop order lottery, but I don't believe that would be better for chasers. If we assume that there are 50 chasers all chasing 10 domains on equal footing, the drop order means you're competing with an average of 5 others for the first domain, with the number of competitors increasing as the list shortens, with all 50 of you finally chasing the last to drop. With the drop order published, all 50 of you are chasing the same domains at the same time. Instead of having a 20% chance lengthening to 2%, every domain is 2%.

Why would all registrars involved be on an equal footing? The vast majority of registrars involved in drop catching today are probably facilitated by a much smaller number of developers. If those developers could capture the vast majority of domain names themselves it reduces the number of registrars who will be successful to those that can develop their own solution. When Nominet created the DAC they probably didn't envisage the concept of DAC hosts as has now become commonplace and likely presumed everyone involved would develop their own solutions and compete solely on technical ability. If developers, coding for a hypothetical drop cycle with published domain drop times so no need to use the DAC, decide to share their code in some manner they'd just be reducing their chances of registering the small pool of contested domain names each day themselves for no real upside. Why do that when you're good enough to develop something and use it with just your own registrar/EPP Create daily total to register many more domain names than ever before? Everyone who can't develop as well can see you at the auction house.

Summarisation: those with the development skills might be able to do significantly better than ever with just one registrar "tag" and everyone who wasn't as good would be using the auctions, placing retail backorders with registrars that were able to compete or out of the business.

 

[RobM]

Will such hypothetical "hundreds of registrars ('tags')" come with the technical ability to develop a system to be able to successfully register very many/any contended domain names at their published drop times?

I think you missed the point. The amount of people entering wouldn't increase - just the tags. If each tag gives you a 1% increased chance, for example, then people who multitag would weigh up the cost of new tags against the increased chance rate and likely selling price of domains.
Anyway this is all hypothetical - we have no idea what nominet are going to do. Just they have *never* introduced anything that has been beneficial for anyone except themselves and the top 5 tags.

 

[BMW7series]

I think you missed the point. The amount of people entering wouldn't increase - just the tags. If each tag gives you a 1% increased chance, for example, then people who multitag would weigh up the cost of new tags against the increased chance rate and likely selling price of domains.
Anyway this is all hypothetical - we have no idea what nominet are going to do. Just they have *never* introduced anything that has been beneficial for anyone except themselves and the top 5 tags.

These registrar 'tags' all have to have unique operators, entities or persons fronting them and someone to whom any domain names they catch are registered. They also need to be financed and managed. They can't all belong to the same person or entity because if they did I believe their EPP create quotas would be linked. It was only with the ROR release where quotas were not linked across credit accounts. In addition many such registrars require someone to provide them with a hosted drop catching system to even be able to operate as drop catchers because they are incapable of producing this sort of system them self. At the moment DAC hosts provide that for a great many registrars.

I accept there is an open ended debate about whether exponentially more non-linked registrar 'tags' providing more EPP Creates would really introduce much of a gain beyond just one, based on the assumption of a drop cycle involving unique drop times for each domain name being published. If only a single domain name drops at any given timestamp and if one has developed a new script which can deal with the time syncing and time offsets better than most others I am unsure how sending many more EPP Creates would be much of a benefit.

I also don't believe there is already a firm decision of an outcome. The idea of publishing drop lists with exact and unique domain name drop times seems to negate the requirement for many to make a significant number of DNS queries and DAC queries for both accurate list building and domain name drop catching. The live DAC might no longer need to be maintained and the complaints about technical flaws with its operation in respect of domain name drop catching would probably cease. The delayed DAC could be maintained for retail register web site "name spinners" and other operations.

Given I don't believe there is already a firm decided outcome, the consultation that is currently open affords anyone the opportunity of putting together alternative suggestions. That could even be to leave things as they are. Of course one needs to be able to properly justify and explain ones views and opinions. One can't simply reply with "leave it be" and expect to be taken seriously! Other *hypothetical* options that I am pulling out of the air might include, for example, publishing approximate drop times instead of specific ones thus still requiring some level of checking, suggesting Nominet switch to using the EPP check command rather than the DAC, limiting participation in drop catching to registrars that sign up to using a special EPP hostname (similar to the ROR) and potentially making that chargeable per annum or per month (flat fee, no tiered access) replacing the current annual fee for the DAC. What might a reasonable monthly fee, for using a drop catching only EPP hostname which might make multi tagging more unattractive, be? Auctions are clearly out of the question. I believe well thought out responses will be taken seriously. If you don't, okay.

 

[Siusaidh]

Auctions are clearly out of the question

I have appreciated the careful analysis you have presented (of course, the truth is, we are all speculating about possible outcomes at this point).

But I am curious to know the careful analytical basis for your statement about auctions.

All auctions do is sell domains like Nominet does already, but with the option to create rational order from the present chaotic rush to catch drops.

Don't get me wrong. I wouldn't benefit from Nominet-based auctions, because I would usually be outbid. But wouldn't a Nominet auction system (triggered by a minimum £50 bid for a name on a pre-published list) be far simpler and more open than the present pressure of the dropcatching system?

It would basically remove drop-catching altogether, which would reduce pressure and strain from the system. The auction system could be operated by two people in a garden shed.

Simplicity?

Fair and open access to every name?

I see no signs that Nominet intend to pursue this concept. I'm curious to understand your reasons why it is 'clearly out of the question'?

 

[ian]

I also don't believe there is already a firm decision of an outcome. The idea of publishing drop lists with exact and unique domain name drop times seems to negate the requirement for many to make a significant number of DNS queries and DAC queries for both accurate list building and domain name drop catching. The live DAC might no longer need to be maintained and the complaints about technical flaws with its operation in respect of domain name drop catching would probably cease. The delayed DAC could be maintained for retail register web site "name spinners" and other operations.

Given I don't believe there is already a firm decided outcome, the consultation that is currently open affords anyone the opportunity of putting together alternative suggestions. That could even be to leave things as they are. Of course one needs to be able to properly justify and explain ones views and opinions. One can't simply reply with "leave it be" and expect to be taken seriously! Other *hypothetical* options that I am pulling out of the air might include, for example, publishing approximate drop times instead of specific ones thus still requiring some level of checking, suggesting Nominet switch to using the EPP check command rather than the DAC, limiting participation in drop catching to registrars that sign up to using a special EPP hostname (similar to the ROR) and potentially making that chargeable per annum or per month (flat fee, no tiered access) replacing the current annual fee for the DAC. What might a reasonable monthly fee, for using a drop catching only EPP hostname which might make multi tagging more unattractive, be? Auctions are clearly out of the question. I believe well thought out responses will be taken seriously. If you don't, okay.

David, is that you?

The DAC is hardly overloaded; 432,000 queries per tag, a few hundred tags, is hardly ddos levels and don't forget Nominet charge for DAC usage. If the purpose of this consultation is to fall in line with .com practices, then fine, but we all know who benefit from .com drops, and it isn't the little guys.

 

[Siusaidh]

Another question. Webinars are often saved and made available for later viewing online. I was at a funeral yesterday afternoon, and so couldn't listen to the webinar. Does Nominet provide the option of watching the webinar later?

 

[ian]

Another question. Webinars are often saved and made available for later viewing online. I was at a funeral yesterday afternoon, and so couldn't listen to the webinar. Does Nominet provide the option of watching the webinar later?

Yeah, drop Leanne an email to [email protected] for the password, as it wouldn't be appropriate for me to give it. The link is here, https://vimeo.com/showcase/5366830

 

[Siusaidh]

Thanks Ian.

 

[RobM]

They can't all belong to the same person or entity because if they did I believe their EPP create quotas would be linked.

Oh that's alright - won't happen then.

 

[davedevelopment]

If only a single domain name drops at any given timestamp and if one has developed a new script which can deal with the time syncing and time offsets better than most others I am unsure how sending many more EPP Creates would be much of a benefit.

I think that with whatever load balancers and/or message queues that are between us and the EPP servers, it's not the first past the post, it's the first past the second or third post. So I think dropcatchers will continue to fire multiple EPP requests.

 

[RobM]

Nominet load balancers struggle to cope with the EPP requests as it is... imagine how they're going to be if everyone is hitting it at the same time. I think you're wasting your time though Dave - it's clear from BMW posts that he doesn't really know how dropcatching works and what effect this hypothetical change would have and I don't think he's here to learn anything. We'll just have to wait and see.

 

[Aaron Clifford]

If the release becomes similar to ROR, the barrier to entry for chasing would become much lower. Right now, you need to perform a DAC check, get a response, and send an EPP request to register the domain, while competing with numerous others doing the same thing. That requires a very fast connection as close as possible to Nominet's servers, and a fast server to actually process the request in time. A known drop time means you simply need a consistent response time and a bit of trial and error to determine the optimal time to make your registration attempt. A $5 Digital Ocean droplet could achieve this easily.

Of course, the actual catch is only a small part of operating a catch service. If you're not getting enough catches because of the amount of competition, the administration would quickly become more trouble than it's worth.

This isn't accurate at all by any stretch, there is more to drop catching nowadays then just checking the DAC, getting a response and sending an EPP request. A known drop time is nothing like RoR, during RoR you had thousands and thousands of names all dropping at the same time, so you may have 50/100 tags going for a particular name, with the presumed setup of every domain having a drop time associated to it, you will have 500 tags all going for that exact domain name at the same time.

 

[stitchbob]

It doesn't matter whether you're dropping 10,000 domains at 2pm or 100 domains every half hour at pre-published times. My point is that without the requirement for a DAC check, you're no longer as reliant on the speed of your connection, since you're only sending one CREATE command.

Of course you're still going to be competing with everybody else, but the actual process of registering a domain is much simpler.