Hi, the story of my latest problem is a bit long, but to me it is serious enough to be worth this explanation.
I observed that since a few days, maybe a week or so, the quality of my gnet connection is suddenly much lower than before. If this is a problem with a recent change in gtkg, this would be bad and should be fixed before 0.95. Perhaps there is another reason for this, and everything is fine except for me, but I'm sceptical. First let me explain what I mean by "quality". I'm not talking about a bandwidth problem or any other problem that could be induced by my ISP. After a year or so of searching for a good quality measure of a given UP connection (I am always a leaf node), I finally came back to the very simple [QHit(TX) per minute], e.g. the rate of query hits I send to that UP. This has proven to me to be a sensible criterion, and nowadays it is no longer distorted by bad clients flooding the net with repeated queries. I know that the absolute value of this measure is not comparable between you and me, because it strongly depends on what I share, but it is a good relative measure to compare the quality of my connections at any given time, and even between different times provided my library of shared media doesn't change. And now, since some days, the average QHit(TX)/min has seriously dropped (my shared stuff has not changed, of course). I remember that only two weeks ago, it was easy to fill up my four slots with UPs that have at least 4 QHit(TX)/min on average each, and now it's even hard to get 2 QHit(TX)/min each. Remember: Don't compare these absolute values to yours, only think of an _at_least_ 50% drop within a week or so. Since this applies to both LimeWire and BearShare UP connections, it is unlikely that this is due to a change in another servent's behaviour. Now, does this decreased gnet "quality" really lead to a decreased upload throughput? Yes, it does. I normally have an average upload of about 10 MB/h, and now it is, as a rough estimation, at least 80% lower!!! This hurts. Unfortunately, I cannot tell why this happens. I don't know if my host receives less queries than before, because I don't know how many queries I received before. And _if_ I do indeed receive fewer queries, I wouldn't know why. Maybe there was a change in the QRP implementation? (only a quick guess, I'm not so deeply into the protocol). Above all, something seems to be broken. Can anyone confirm my observations? Greetz, Hauke Hachmann ------------------------------------------------------- This SF.Net email is sponsored by: InterSystems CACHE FREE OODBMS DOWNLOAD - A multidimensional database that combines robust object and relational technologies, making it a perfect match for Java, C++,COM, XML, ODBC and JDBC. www.intersystems.com/match8 _______________________________________________ Gtk-gnutella-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/gtk-gnutella-devel
