New domain expiry process and droplists for .uk

[JamesTuplin]

And under the new system additional tags would be at cost WITH or without additional creates?

 

[dee]

In the current system there is a real benefit in getting additional memberships and tags, despite being clearly against the rules.
In the proposed system additional memberships will be meaningless since you can get additional tags on your account. But I must stress that Nominet believes additional tags will not make a difference, it's all about timing.

Yes. Think i get the idea ... but that was my other point. If it's all about timing, the surely the person that has the one hop server and is first to the exact timing / latency send will clear up ? Or the person that is very close to exact timing and has two sets of sends seperated by by half a millisecond will clear up ....Or the person that has 3 ..... etc etc

Maybe im not getting it.

 

[dee]

And under the new system additional tags would be at cost WITH or without additional creates?

if you are declaring them as connected . That was always one of the issues wasnt it ?

 

[dee]

I'm either am not understanding it correctly or this really doesn't make sense to me?

So 'Mr X' & Mr Y' both get there latency and timing spot on, however 'Mr X' is only using 1 TAG with 6 EPP Sockets to fire off EPP CREATE requests, but 'Mr Y' is utilising 3 TAGS and 18 EPP Sockets to fire off EPP CREATE requests - surely 'Mr Y' has an advantage, similar to him having 3 raffle tickets to 'Mr X's' 1?

As you well know, I am not the most technical, so I may well have interpreted this all wrong, but this is my current understanding of the proposal.

this.... you put it better than i did.

 

[webber]

I am a member and I wish for my parents, grandparents, children, kids and the milkman to become members (each with their own quota). Would this be ok?

Nothing changes in regards to membership or tags.
You still need to declare connected tags, these still share any quotas, etc.
Nominet are only implementing droplists, which was overwhelmingly supported in the 2019 consultation.
Hence why the UKRAC work was called DLIWG (Drop Lists Implementation Working Group) and not something else.

'Mr X' is only using 1 TAG with 6 EPP Sockets to fire off EPP CREATE requests, but 'Mr Y' is utilising 3 TAGS and 18 EPP Sockets to fire off EPP CREATE requests - surely 'Mr Y' has an advantage, similar to him having 3 raffle tickets to 'Mr X's' 1?

Not necessarily. It really depends on how good their timing is.
Sounds to me like Mr X has a very good solution and he is confident his timing is spot on, while Mr Y is kind of winging it and firing as many <create> as possible in the hope he might get lucky. If Mr X's timing is indeed spot on then he can still catch the domain with a single <create>
We'll have to wait and see I suppose. But if additional tags is a proven benefit then anyone is free to sign up for an additional tag and this will all be transparent to everyone else.

 

[dee]

Not necessarily. It really depends on how good their timing is.
Sounds to me like Mr X has a very good solution and his timing is spot on, while Mr Y is kind of winging it and firing as many <create> as possible in the hope he might get lucky. If Mr X's timing is indeed spot on then he can still catch the domain with a single <create>
We'll have to wait and see I suppose. But if additional tags is a proven benefit then anyone is free to sign up for an additional tag and this will all be transparent to everyone else.

Out of interest. What is the timing resolution for this stuff?

And... you're saying once the timing is right then its a done deal ? It wont take long to work out the timing for the tech bods, and then its just repeat the fire at correct interval and win then ?

Not trying to be facetious, just trying to understand.

 

[webber]

And... you're saying once the timing is right then its a done deal ? It wont take long to work out the timing for the tech bods, and then its just repeat the fire at correct interval and win then ?

IP data takes different routes at different times even if nothing else changes.
Most likely your software will have to determine the latency every time it connects, constantly measure it, and then compensate for that. Maybe even factor in any additional network load that might be expected at that specific moment when it sends the <create> if it is a particularly popular domain.

It will really be a technical race. Some members will love this and tinker with their software and constantly improve it. This is also the case now. Other members will wing it and flood the EPP with 1000 <create> commands. Some members can do both.

Let's be honest, currently there are a few tags, using highly tuned software and servers, that we expect to catch a highly disputed domain in a day. There is an element of randomness to the whole thing so one can expect to get lucky once in a while if they use a decent enough software and server, and maybe a more narrowly focused strategy.
I think this will continue to remain mostly the same – there will be those that can dedicate the time to write great software and tweak it every day; they will always reap the rewords for their work. But I think there will also be an element of luck for those firing their entire 1000 <create> quota on a single domain.
We'll just have to work out what works and what doesn't, as we have done with the DAC.

 

[dee]

IP data takes different routes at different times even if nothing else changes.
Most likely your software will have to determine the latency every time it connects, constantly measure it, and then compensate for that. Maybe even factor in any additional network load that might be expected at that specific moment when it sends the <create> if it is a particularly popular domain.

It will really be a technical race. Some members will love this and tinker with their software and constantly improve it. This is also the case now. Other members will wing it and flood the EPP with 1000 <create> commands. Some members can do both.

Let's be honest, currently there are a few tags, using highly tuned software and servers, that we expect to catch a highly disputed domain in a day. There is an element of randomness to the whole thing so one can expect to get lucky once in a while if they use a decent enough software and server, and maybe a more narrowly focused strategy.
I think this will continue to remain mostly the same. There will be those that can dedicate the time to write great software and tweak it every day. Those will always reap the rewords. But there will also be an element of luck for those firing their entire 1000 <create> quota on a single domain.

Yes. Again though.... theres advantage to be had being in close proximity and a short hop away from reciving server. Thats always been the case.

 

[webber]

theres advantage to be had being in close proximity and a short hop away from reciving server. Thats always been the case.

That's definitely the case now. It will no longer be the case, for sure. See this:

Theoretically, if I was in Australia I could send my request off 200 ms ahead of time and still beat someone who is 0.5 ms away if I timed it right.

 

[lazarus]

And... you're saying once the timing is right then its a done deal ? It wont take long to work out the timing for the tech bods, and then its just repeat the fire at correct interval and win then ?

Too many variables to set and forget. It will have to be constantly tweaked.

 

[dee]

That's definitely the case now. It will no longer be the case, for sure. See this:

Again, i may just not be getting this, but the traceroute from Australia is gonna be zipping around all over the place, through many routes,compounding errors and timing. Thats different to having a server in the same building isnt it ??

 

[dee]

Too many variables to set and forget. It will have to be constantly tweaked.

okay. I guess im not understanding the tech side. Guess thats a good thing. Although also a good reason to have multiple tags with different settings maybe .. possibly ?

 

[lazarus]

Again, i may just not be getting this, but the traceroute from Australia is gonna be zipping around all over the place, through many routes,compounding errors and timing. Thats different to having a server in the same building isnt it ??

If you know the time it takes to get from point A to point B then you can just send the request that much ahead of time.

 

[webber]

Again, i may just not be getting this, but the traceroute from Australia is gonna be zipping around all over the place, through many routes,compounding errors and timing. Thats different to having a server in the same building isnt it ??

Depends how many hops your route takes. You could be plugged into a tier1 network in Australia and be just a few hops away, albeit still 200ms latency. I wouldn't recommend it, but it could/should work

 

[dee]

If you know the time it takes to get from point A to point B then you can just send the request that much ahead of time.

But doesn't change more dramatically every time the further away you are ? If your one hop its a lot more consistent

I know you guys are a lot more capabe than me on the programming side. I'll get my coat.

 

[lazarus]

But doesn't change more dramatically every time the further away you are ? If your one hop its a lot more consistent

Can do I guess but thats an extreme example. It would still be good to be close as possible I guess. In the real world the difference from Mythic Beasts or Digital Ocean may be only 1ms or less so you could tweak it based on your own specifics at that time.

 

[JamesTuplin]

Nominet are only implementing droplists, which was overwhelmingly supported in the 2019 consultation.
Hence why the UKRAC work was called DLIWG (Drop Lists Implementation Working Group) and not something else.

But we didn't overwhelming support drop lists with EXACT drop times meaning complete re-writes to drop catching systems. I cannot ever remember agreeing to this, nor would I ever have agreed to it. I cannot see how publishing exact drop times is going to make this situation better for anyone. Unless Nominet can tell me how greater competition (more people catching domains) is better for my business. There's more incentive now than ever to become a member and collude.

Perhaps if more time had been spent to clarify collusion. Is it:

• 1 member with multiple sub members (family members, friends, milkman) catching domains to benefit a single person?
• Public drop catchers using your DAC allowance to query domains for others without your permission?
• Forced to sell domain(s). Sign up for platform 123, but you cannot keep the domain, you must sell it?

And then re-write the terms and conditions with clearly defined repercussions to deter others from doing so.

 

[webber]

But we didn't overwhelming support drop lists with EXACT drop times meaning complete re-writes to drop catching systems. I cannot ever remember agreeing to this, nor would I ever have agreed to it.

Maybe you didn't or you don't remember, but 89% of respondents did.
If you have any issues with how that consultation was conducted, let Nominet know.
I didn't post about this here to start a heated discussion or to defend Nominet, I just wanted to inform those who maybe didn't receive the notification because they opted out of email communication from Nominet.
I highly recommend you enable those because Nominet will not knock on your door to ask for your opinion. You will get an email like everyone else, if you haven't opt out. GDPR and all...

Personally I think this new system will actually work in your favour and anyone who catches a large number of domains each day. You will no longer have to split your fixed daily quota over the number of domains you wish to catch. You should have an equal change at catching a highly disputed domain AND 20 or 30 undisputed SEO/traffic domains.

 

[JamesTuplin]

if you haven't opt out. GDPR and all...

Like: https://www.acorndomains.co.uk/threads/nominet-uk-rdap.173119/ you mean?

Personally I think this new system will actually work in your favour and anyone who catches a large number of domains each day. You will no longer have to split your fixed daily quota over the number of domains you wish to catch. You should have an equal change at catching a highly disputed domain AND 20 or 30 undisputed SEO/traffic domains.

Sounds like a typical Nominet response. Im sorry. I know you are doing your best here @webber and to be fair I commend you for being the only person to keep everyone updated, but I feel like we have been sold down the river. Time will prove that I am right.

 

[JamesTuplin]

Forgive me if this is obvious, but how will 20/30 SEO domains be undisputed? The only way we'd be looking at winning anything here is if several high-value names drop at the same minute. How often will that happen? Otherwise Mr Fast with 90 Nominet tags at 2:15 will get LL.co.uk and at 2:16 will get reallygoodseodomain.co.uk etc.

Anyone that thinks otherwise needs to shake themselves silly and sit the corner of the room, alone.