- Joined
- Dec 17, 2009
- Messages
- 151
- Reaction score
- 0
As an ISP, thought we might drop a line or two on our development and implementation of an Enom Drop Catch solution, and what we found on the way.Having been an ISP since 93, we, as many others, constantly kick ourselves on the names we could have had, and the ones we let go, many at cost without realising worth, still we plodded on , and at least we are still here, so that's one good thing..
So, with the onset of Nominets DAC and EPP provisions, about 2 years ago we set about developing our own EPP / DAC script with a small team, not such a hard job you may think, but with the essence being time slots, it became a task concentrating solely on microsecond reductions, with the end result being a total DAC and EPP script with only a page or so of code, and for all that work !
Testing the EPP portion became expensive in realtime, as in most cases the developers didn't use the test platform, but with every run, registered names... most of them meaningless, but at least it finally became ultra fast, within micro seconds. After constant tweaking over 10 months, the software was bundled and released for sale, which it still is, and gets relatively moderate and consistent sales, all very good you may say, but what about results...
Well we know the software works extremely well, from DAC to EPP in microseconds, including the keepalive needed by the EPP server, but where are all the dropped name registrations ? mmm.... we took a closer look at this... and whilst the drop catching was pretty good as far as relatively known domains with mediocre popularity were caught, the big releases were not.. so why was this, and why on investigation was it always the same three or four people catching them. how can that be when we all have to adhere to Nominet's policy's ?
Suppose there are some obvious answers to this, one being that maybe our servers were not close enough to the Nominet rack in Telehouse, thereby our microsecond DAC to EPP, although unbeatable (we think), propagation times were hindering us. So, we obtained a shell on a clear server the same AS as Nominet, mmm... not much difference there, so whats going on ?Nominets DAC is limited, so as a tag holder with a DAC account, restricts you to a certain number of automated lookups per 24 hour period, any more and your blocked, this equates to around 180 ms per lookup if your running the script for 24 hours, and in propogation terms, thats a lot, so for instance if your DAC lookup gets a released response and drops to the EPP in a flash, it may be way way too late, in fact it may be 180ms too late..
How can we get rid of this huge delay, well, have 180 Nominet tags, mm, bit expensive (at £100 a year each) and probably Nominet will have a word or two to say about this, never the less, if the script is delayed by 1ms, its feasable.Or what about just running the script for 6 hours a day, that gives you 180/4 = 45ms , much better, although you risk a chance of trying to catch a drop that's already been registered..So there it is, impeccable software, but a quandry when it comes to aquisition, but it still feels like winning the lottery when you do catch a goody !
CheersBill
So, with the onset of Nominets DAC and EPP provisions, about 2 years ago we set about developing our own EPP / DAC script with a small team, not such a hard job you may think, but with the essence being time slots, it became a task concentrating solely on microsecond reductions, with the end result being a total DAC and EPP script with only a page or so of code, and for all that work !
Testing the EPP portion became expensive in realtime, as in most cases the developers didn't use the test platform, but with every run, registered names... most of them meaningless, but at least it finally became ultra fast, within micro seconds. After constant tweaking over 10 months, the software was bundled and released for sale, which it still is, and gets relatively moderate and consistent sales, all very good you may say, but what about results...
Well we know the software works extremely well, from DAC to EPP in microseconds, including the keepalive needed by the EPP server, but where are all the dropped name registrations ? mmm.... we took a closer look at this... and whilst the drop catching was pretty good as far as relatively known domains with mediocre popularity were caught, the big releases were not.. so why was this, and why on investigation was it always the same three or four people catching them. how can that be when we all have to adhere to Nominet's policy's ?
Suppose there are some obvious answers to this, one being that maybe our servers were not close enough to the Nominet rack in Telehouse, thereby our microsecond DAC to EPP, although unbeatable (we think), propagation times were hindering us. So, we obtained a shell on a clear server the same AS as Nominet, mmm... not much difference there, so whats going on ?Nominets DAC is limited, so as a tag holder with a DAC account, restricts you to a certain number of automated lookups per 24 hour period, any more and your blocked, this equates to around 180 ms per lookup if your running the script for 24 hours, and in propogation terms, thats a lot, so for instance if your DAC lookup gets a released response and drops to the EPP in a flash, it may be way way too late, in fact it may be 180ms too late..
How can we get rid of this huge delay, well, have 180 Nominet tags, mm, bit expensive (at £100 a year each) and probably Nominet will have a word or two to say about this, never the less, if the script is delayed by 1ms, its feasable.Or what about just running the script for 6 hours a day, that gives you 180/4 = 45ms , much better, although you risk a chance of trying to catch a drop that's already been registered..So there it is, impeccable software, but a quandry when it comes to aquisition, but it still feels like winning the lottery when you do catch a goody !
CheersBill
Last edited:
