Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!uakari.primate.wisc.edu!news.arc.nasa.gov!dftsrv!heawk1!hoepfner From: hoepfner@heawk1.gsfc.nasa.gov (Patrick Hoepfner) Newsgroups: comp.sys.mac.comm Subject: Re: Downloading with Kermit (not slow with packet size 1000!) Message-ID: Date: 17 Jun 91 04:15:34 GMT References: <1991Jun7.112254.622@vax.oxford.ac.uk> <1991Jun12.205728.997@m.cs.uiuc.edu> Sender: news@dftsrv.gsfc.nasa.gov Organization: NASA/Goddard Space Flight Center Lines: 29 ingemar@isy.liu.se (Ingemar Ragnemalm) writes: >gillies@m.cs.uiuc.edu (Don Gillies) writes: >>kermit is perhaps the slowest download protocol in existence. It is >>slow because the packet sizes are too small (128 bytes). If you use a >>better protocol (xmodem, ymodem, or especially zmodem) you would be >>much better off. In fact, binhex + zmodem would probably run faster >>than macbinary + kermit. >I tried Zmodem protocol (with Zterm), and for some reason it was slower than >MacKermit. This was because of two reasons: >- The Zmodem protocol didn't seem to work perfectly. It got some errors in >transmission, which is strange since Kermit had no problems. >- I used packet sizes over 1000 for MacKermit. This is much faster than the >standard packet length. If you use a comercial program (WhiteKnight for example) you will find that ZModem is still faster that Kermit. ZModem is a streaming protocol. Even using SuperKermit on your host (the streaming version of Kermit) ZModem is faster. It uses 1024 byte packets and you don't have the overhead that Kermit has for binary files. I don't have the streaming version of Kermit, I can't seem to find it for our Sun or Vax. -- Pat --------------------------------------> hoepfner@heasfs.gsfc.nasa.gov