RE: IP->Country Data (RE: ISP's Contact List)
I'm sure that this made it here before but: www.analysespider.com/ip2country/ip_country.html Later, J -----Original Message----- From: william(at)elan.net [mailto:william@elan.net] Sent: Monday, June 13, 2005 1:40 PM To: nanog@merit.edu Subject: IP->Country Data (RE: ISP's Contact List) On Mon, 13 Jun 2005, Daniel Senie wrote:
had wrong info or no info.
If you're the ISP, then your answer to your customers is "the web site
you are going to is using an unreliable method to attempt to determine
locality from IP address."
Then point them at the web site's contact form and suggest they complain to the web site owner.
As you've already noted, ARIN got the info correct. I suspect there are multiple vendors attempting to provide IP address to GPS address services. I doubt any of them will be 100% reliable, but clearly there
is an interest in such services.
FYI - It maybe of interest here to know that completewhois provides this data. For determining ip->country placement only RIR data on direct ip allocations and assignments is used, so if ip block is listed as having been allocated to company with corporate address in Jamaica but is used in US, the block will show up as JM. That also avoids issue of having others blame me for bad data as algorithm for how its generated is easily confirmed with RIR whois and in 98% this does provide sufficiently good data. Our lists are regenerated EVERY DAY and available at: http://www.completewhois.com/statistics/data/ips-bycountry/rirstats/ Even though I've never promoted this, it actually is getting rather used and I've had at least dozen requests this year along to make data available though means other then raw text file. So you might as well be first to know that 3 days ago this was finished and data is now available for verification by dns. You can now do lookup in RBL style to <reverse-ip>.country-rirdata.dnsiplists.completewhois.com (i.e. for my 216.151.192.1 dns server ip, it would be lookup at 1.192.151.216.coutry-rirdata.dnsiplists.completewhois.com) For TXT lookups it will tell you country code and country name, i.e. "US - United States". For RBL "A" lookups it will answer with 127.0.a.b where a and b are ascii representation of 1st and 2nd letter of country code, so for example for US this would be 127.0.85.89 Complete list of these codes and instructions are at: http://www.completewhois.com/statistics/data/ips-bycountry/rirstats/READ ME So feel free to use if you like for whatever reasons (and don't complain to me if you see wrong data, lookup RIR whois and send corrections there). One last point is as I noted data is based on RIR direct allocations, so swips and suballocations are not used. In the future planned is separate service and data for experimental use that will be based on SWIP data (lowest allocation), however my checks on it show that in too many places its even less accurate then RIR direct info. -- William Leibzon Elan Networks william@elan.net
On Mon, 13 Jun 2005, Jim McBurnett wrote:
I'm sure that this made it here before but:
www.analysespider.com/ip2country/ip_country.html
Commercial service when there are several free ones available.... BTW - based on what I can see they are not updating data once per day but only once/month. Also having some experience in this matter I highly doubt that its much more accurate then what I got - it appears no matter what path you take to get the data, it would to a degree be wrong (difference being which part is wrong). One possible way to resolve some issues could be to have testing servers placed in several locations and analyze TTL for connections to various ips to determine how far its from several servers - of course it would also not be accurate as ip networks are not necessarily interconnecting in the same region (CAIDA did their best with this method some time ago but have not kept updated data as far as I know).
-----Original Message----- From: william(at)elan.net [mailto:william@elan.net] Sent: Monday, June 13, 2005 1:40 PM To: nanog@merit.edu Subject: IP->Country Data (RE: ISP's Contact List)
On Mon, 13 Jun 2005, Daniel Senie wrote:
had wrong info or no info.
If you're the ISP, then your answer to your customers is "the web site
you are going to is using an unreliable method to attempt to determine
locality from IP address."
Then point them at the web site's contact form and suggest they complain to the web site owner.
As you've already noted, ARIN got the info correct. I suspect there are multiple vendors attempting to provide IP address to GPS address services. I doubt any of them will be 100% reliable, but clearly there
is an interest in such services.
FYI - It maybe of interest here to know that completewhois provides this data. For determining ip->country placement only RIR data on direct ip allocations and assignments is used, so if ip block is listed as having been allocated to company with corporate address in Jamaica but is used in US, the block will show up as JM. That also avoids issue of having others blame me for bad data as algorithm for how its generated is easily confirmed with RIR whois and in 98% this does provide sufficiently good data. Our lists are regenerated EVERY DAY and available at: http://www.completewhois.com/statistics/data/ips-bycountry/rirstats/
Even though I've never promoted this, it actually is getting rather used and I've had at least dozen requests this year along to make data available though means other then raw text file. So you might as well be first to know that 3 days ago this was finished and data is now available for verification by dns. You can now do lookup in RBL style to <reverse-ip>.country-rirdata.dnsiplists.completewhois.com
(i.e. for my 216.151.192.1 dns server ip, it would be lookup at 1.192.151.216.coutry-rirdata.dnsiplists.completewhois.com)
For TXT lookups it will tell you country code and country name, i.e. "US - United States". For RBL "A" lookups it will answer with 127.0.a.b where a and b are ascii representation of 1st and 2nd letter of country code, so for example for US this would be 127.0.85.89
Complete list of these codes and instructions are at:
http://www.completewhois.com/statistics/data/ips-bycountry/rirstats/READ ME
So feel free to use if you like for whatever reasons (and don't complain to me if you see wrong data, lookup RIR whois and send corrections there).
One last point is as I noted data is based on RIR direct allocations, so swips and suballocations are not used. In the future planned is separate service and data for experimental use that will be based on SWIP data (lowest allocation), however my checks on it show that in too many places its even less accurate then RIR direct info.
-- William Leibzon Elan Networks william@elan.net
----- Original Message ----- From: "william(at)elan.net" <william@elan.net> To: "Jim McBurnett" <jim@tgasolutions.com> Cc: <nanog@merit.edu> Sent: Monday, June 13, 2005 2:55 PM Subject: RE: IP->Country Data (RE: ISP's Contact List)
On Mon, 13 Jun 2005, Jim McBurnett wrote:
I'm sure that this made it here before but:
www.analysespider.com/ip2country/ip_country.html
Commercial service when there are several free ones available....
BTW - based on what I can see they are not updating data once per day but only once/month. Also having some experience in this matter I highly doubt that its much more accurate then what I got - it appears no matter what path you take to get the data, it would to a degree be wrong (difference being which part is wrong). One possible way to resolve some issues could be to have testing servers placed in several locations and analyze TTL for connections to various ips to determine how far its from several servers - of course it would also not be accurate as ip networks are not necessarily interconnecting in the same region (CAIDA did their best with this method some time ago but have not kept updated data as far as I know).
the most accurate systems i am aware of use online purchase statistical data to increase accuracy. someone purchases a blue widget online, from a certain ip address and with a credit card with a certain billing zip code, and voila - they now have an idea of where that ip might be. of course, this works best (only?) for residential ip addresses in broadband-enabled neighbourhoods, but it's pretty darn good at telling me where i am (within 5-10 miles). -p
On Jun 13, 2005, at 2:55 PM, william(at)elan.net wrote:
Commercial service when there are several free ones available....
BTW - based on what I can see they are not updating data once per day but only once/month. Also having some experience in this matter I highly doubt that its much more accurate then what I got - it appears no matter what path you take to get the data, it would to a degree be wrong (difference being which part is wrong). One possible way to resolve some issues could be to have testing servers placed in several locations and analyze TTL for connections to various ips to determine how far its from several servers - of course it would also not be accurate as ip networks are not necessarily interconnecting in the same region (CAIDA did their best with this method some time ago but have not kept updated data as far as I know).
Too bad no one has servers in 100s or even a 1000 or more ISPs in dozens of countries with TCP connections to a statistically significant portion of the Internet on a daily basis who could possibly measure, say, RTT (not sure why TTL is relevant) and other things, and perhaps use that along with other information to create a database that they could then give to customers who need geo-location data for their "mission critical" applications. That would probably be pretty darned accurate. Of course, that would be a "commercial service".... =) -- TTFN, patrick
At 6:35 PM -0400 2005-06-13, Patrick W. Gilmore wrote:
Too bad no one has servers in 100s or even a 1000 or more ISPs in dozens of countries with TCP connections to a statistically significant portion of the Internet on a daily basis who could possibly measure, say, RTT (not sure why TTL is relevant) and other things, and perhaps use that along with other information to create a database that they could then give to customers who need geo-location data for their "mission critical" applications.
That would probably be pretty darned accurate.
You mean like Akamai?
Of course, that would be a "commercial service"....
You mean like Akamai?
=)
;-) -- Brad Knowles, <brad@stop.mail-abuse.org> "Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety." -- Benjamin Franklin (1706-1790), reply of the Pennsylvania Assembly to the Governor, November 11, 1755 SAGE member since 1995. See <http://www.sage.org/> for more info.
Too bad no one has servers in 100s or even a 1000 or more ISPs in dozens of countries with TCP connections to a statistically significant portion of the Internet on a daily basis who could possibly measure, say, RTT (not sure why TTL is relevant) and other things, and perhaps use that along with other information to create a database that they could then give to customers who need geo-location data for their "mission critical" applications.
Unless I'm mistaken you do (or at least as close to it as anygone got) and that is one of the basis of your service!
That would probably be pretty darned accurate.
The point is it probably would not be as accurate as one could hope because internet network infrastructure is network-centric and not necessarily region-based. Of course you could try to fully map INET like CAIDA does and keep the info updated on "daily basis", pretty hard work though...
Of course, that would be a "commercial service".... =)
Certainly having to support machine on every large network in every city is rather troublesome otherwise. Of course this is probably way overkill for IP->Country data. -- William Leibzon Elan Networks william@elan.net
On Mon, 13 Jun 2005, william(at)elan.net wrote: The point is it probably would not be as accurate as one could hope because internet network infrastructure is network-centric and not necessarily region-based. Of course you could try to fully map INET like CAIDA does and keep the info updated on "daily basis", pretty hard work though... You miss the point; what they have done is decouple the requirement for correlation to a geographical location in optimizing content delivery; the only thing they seem to care about is logical topographical location. As for the 'pretty hard work' part; they seem to be making money off it; how's your gig going? matto --matt@snark.net------------------------------------------<darwin>< The only thing necessary for the triumph of evil is for good men to do nothing. - Edmund Burke
On Jun 13, 2005, at 6:53 PM, Matt Ghali wrote:
On Mon, 13 Jun 2005, william(at)elan.net wrote:
The point is it probably would not be as accurate as one could hope because internet network infrastructure is network-centric and not necessarily region-based. Of course you could try to fully map INET like CAIDA does and keep the info updated on "daily basis", pretty hard work though...
You miss the point; what they have done is decouple the requirement for correlation to a geographical location in optimizing content delivery; the only thing they seem to care about is logical topographical location.
Well, "they" probably look at things like throughput and packloss, as well as logical topology / RTT. But what do I know? :)
As for the 'pretty hard work' part; they seem to be making money off it; how's your gig going?
Not sure that's a fair comparison, since I didn't think William is doing this for money. (William, can you confirm?) Someone offering "less" service, but for "less" (or no) money seems perfectly reasonable. In fact, it's downright nice. But doubt it would be as accurate. (Which is not a slight against William.) -- TTFN, patrick
As for the 'pretty hard work' part; they seem to be making money off it; how's your gig going?
I primarily make money from consulting and other work not from ISP services which I have not promoted from 2002.
Not sure that's a fair comparison, since I didn't think William is doing this for money.(William, can you confirm?)
Certainly true. All the data is available for free and it will not change. To be fair, I do provide consulting service on how to integrate this and similar data in other applications. Anyone else can do the same. -- William Leibzon Elan Networks william@elan.net
On Mon, 13 Jun 2005, Matt Ghali wrote:
The point is it probably would not be as accurate as one could hope because internet network infrastructure is network-centric and not necessarily region-based. Of course you could try to fully map INET like CAIDA does and keep the info updated on "daily basis", pretty hard work though...
You miss the point; what they have done is decouple the requirement for correlation to a geographical location in optimizing content delivery; the only thing they seem to care about is logical topographical location.
I did not "miss the point". I just pointed that such data would not be same as geographic location. That akamai does not need geo data specifcally for their application setup is another matter entirely. -- William Leibzon Elan Networks william@elan.net
participants (6)
-
Brad Knowles -
Jim McBurnett -
Matt Ghali -
Patrick W. Gilmore -
Paul G -
william(at)elan.net