Zlatin Balevsky <[EMAIL PROTECTED]> writes: > We are often sending 10s of messages in a single packet; most of these > messages are very similar (i.e. QRs). While each message in itself > isn't well compressible because of the nature of the strings it > contains, a packet with 20 messages could be compressed much better, > resulting in reduced bandwith usage for high-traffic nodes. > > This compression will of course happen before encryption and will be > completely transparent to the upper layers of the code.
Compression is good. But I disagree with you about the compressibility of individual messages. Of course, that's what I'm working on now; building a compressed network form of the messages fred sends so that the waste of sending hex strings instead of binary data, and longJavaVariableNames instead of single-byte flags (since there's not that many different variables). We shall see how well my experiment works, but there'll still be room to compress 20 copies of the same message into a smaller unit. [1] Thelema [1] except the messages will be like: \012 (meaning "QueryRejected") \001 (meaning "Connection=sustain") \008 (meaning "UniqueID=") bunch of random bytes of UniqueID (8 of them for the 64-bit ID) So the only thing that'll be the same will be the \012\001\008, which doesn't make for good compression, even if there's 20 of them with random IDs interspersed. -- E-mail: [EMAIL PROTECTED] Raabu and Piisu GPG 1024D/36352AAB fpr:756D F615 B4F3 BFFC 02C7 84B7 D8D7 6ECE 3635 2AAB _______________________________________________ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
