Since i got no answer, i decided to check it out my self with my own little experiment:
using my own subnet at home (2 computers linked through a router), i wrote a script to repeatedly send yql queries from a web page.
i ran that script on the 1st computer until i reached the limit (BTW - it was less than 1000, i got block somewhere around the 200th call).
until the limit was reached i got HTTP 200 response, and all was well.
once i've reached the limit - i got HTTP 500 response over and over again.

a minute later, i ran the exact same script on another computer within the same subnet - and i instantly failed with the same HTTP 500 response.

my conclusion : the IP limitation (1,000 calls per hour) is per subnet, and not per computer.
meaning - it's enough to have a single endpoint in your subnet that reached the limit, to disable the entire network from executing YQL queries for the next hour or so...

any comments?

yahoo seems to be very careless about their services. forum messages are widely ignored, while quality and reliability of the services seem to have been dropping severely, but no one could care less. i was really willing to put an extra last hope, but i'm seriously thinking of other options now.

You have a good point here. I personally would like to be able to get information about how many calls are left, when I make a certain request. Other APIs e.g. the twitter API, send information like that in the response header.

A 2nd note regarding your comment:
- If it is just about getting more calls, why not authorize via OAuth?
- I am surprised that you got blocked after 200 calls only. What kind of call did you send?

Hi, I using oauth and send all my requests to /v1/yql/*, but it&#39;s still the same limitations. I&#39;ve got banned every time after ≈500 queries. So it is even less than stated.<br><br><div class="quote "><div class="quotetop ">QUOTE<cite>(Hans Müller @ 24 Jul 2011 6:50 AM)</cite><blockquote class="quotemain">You have a good point here. I personally would like to be able to get information about how many calls are left, when I make a certain request. Other APIs e.g. the twitter API, send information like that in the response header.<br><br>A 2nd note regarding your comment:<br>- If it is just about getting more calls, why not authorize via OAuth?<br>- I am surprised that you got blocked after 200 calls only. What kind of call did you send?<br><br><br><br><!--quoteo(post=24419:date=Jun 19 2011, 03:32 AM:name=omrigabai@...)--><div class="quote "><div class="quotetop ">QUOTE <cite>(omrigabai@... @ Jun 19 2011, 03:32 AM) <a href="index.php?act=findpost&pid=24419">&lt;{POST_SNAPBACK}&gt;</a></cite><blockquote class="quotemain"><!--quotec-->Since i got no answer, i decided to check it out my self with my own little experiment:<br>using my own subnet at home (2 computers linked through a router), i wrote a script to repeatedly send yql queries from a web page.<br>i ran that script on the 1st computer until i reached the limit (BTW - it was less than 1000, i got block somewhere around the 200th call).<br>until the limit was reached i got HTTP 200 response, and all was well.<br>once i&#39;ve reached the limit - i got HTTP 500 response over and over again.<br><br>a minute later, i ran the exact same script on another computer within the same subnet - and i instantly failed with the same HTTP 500 response.<br><br>my conclusion : the IP limitation (1,000 calls per hour) is per subnet, and not per computer.<br>meaning - it&#39;s enough to have a single endpoint in your subnet that reached the limit, to disable the entire network from executing YQL queries for the next hour or so...<br><br>any comments?<!--QuoteEnd--></blockquote><!--QuoteEEnd--></div></div></blockquote></div></div>