Summary: sql access over 9.6 kbps line

Some time ago I posted a question about sql access over 9.6 kbps lines. I asked: >> >>Did anybody use SQL database access over 9.6 kbit/s line ? >>I mean, using TCP/IP and PPP at the transport level ? >> >>What is the kind of speed you get ? >>Any experience with other baud rates over WAN, like 56..64 kbit/s ? >> >>For comparison: I tried X-protocol over 9.6 kbps, and I can say it worked >>in theory and as an example, but in *practice*, everything is so >>slow that very few people would want to use it every day. >> This is a summary of answers I got: >From: "Paul E. Begley" <uunet!scdipc0.ueci.com!peb> >Date: Wed, 25 May 1994 15:57:31 -0400 (EDT) > >As Yaakov Lerner (SDH-S) so eloquently stated: >:In practice, did anybody try or use SQL database access >:over 9.6 kbit/s line ? I mean, using TCP/IP and PPP as transport level ? >: >:What is the kind of speed you get ? >:Any experience with other baud rates over WAN, like 56..64 kbit/s ? > >We are using a 56K line between Philadelphia and Cork, Ireland. The >response time is OK. We have 4GL apps which utilize some lookup tables >and commit record upon approval to the remote tables. We found >response times in the 1-2 second range for complex queries on small >size tables (thousands of records not tens of thousands, outer joins with >subqueries). > >HOWEVER, you should note there is a difference between line speed and >response time! For a several week period our normal fiber optic link >when to a satillite connection. Response time went into the crapper. >Checking ping we found 64 byte packets took 600ms versus 125ms with >fiber. The vendor guarentees line speed, not response time. Something >to be aware of for future reference. > >Depending on the quantity of data and how you are accessing it, your >mileage will vary. > >:For comparison: I tried X-protocol over 9.6 kbps, and I can say it worked >:in theory and as an example, but in *practice*, everything is so >:slow that very few people would want to use it every day. > >There is a big (BIG) difference between SQL calls and running X apps over a >WAN! > >Hope this helps. >-- >Paul Begley Internet: peb@scdipc0.ueci.com >Raytheon Engineers & Constructors Voice: 215-422-3997 >Philadelphia, PA FAX: 215-422-4095 >From: Subhash Ramachandran <uunet!access.digex.net!subhash> >Subject: Using to access SQL database > >>Yaakov Lerner (SDH-S) (lerner@ecitele.com) wrote: >>> >>> In practice, did anybody try or use SQL database access >>> over 9.6 kbit/s line ? I mean, using TCP/IP and PPP as transport level ? > > We are running an Informix database on an Intergraph server (Unix >System V). The server is running a SLIP product called PPP/SLIP and on the >client side we have a SLIP software from PCTCP. > >>> What is the kind of speed you get ? >>> Any experience with other baud rates over WAN, like 56..64 kbit/s ? > > The client is (in most cases) a notebook computer with a 14.4K modem. >On the server side, there is a 19.2K modem. However, the SLIP software can only >be configured for a max data speed of 9600 baud - all transmissions therefore >take place at that speed. Ping times were of the order of 500-550 ms (in >comparison, ping times over a LAN is in the order of 55 ms). > >>> For comparison: I tried X-protocol over 9.6 kbps, and I can say it worked >>> in theory and as an example, but in *practice*, everything is so >>> slow that very few people would want to use it every day. > > Running X terminals using SLIP is probably not a good idea - the >performance is terrible. That's why we decided to use a client-server >approach rather than use X-terms. > > > We did not see great response times for any of the SQL queries, but >they were good enough for the kind of use this system provided. An important >lesson we learnt was queries need to be optimized if you are to see good >retrieval times. > >Hope this helps. > >Subhash >From: uunet!deepthwt.demon.co.uk!ceisen (Clive Eisen) >Organization: Clive Eisen Computing >Subject: ? sql access over 9.6 kbps line ? > >In article <9405251259.AA13075@ecitele.com> you write: > >> >> In practice, did anybody try or use SQL database access >> over 9.6 kbit/s line ? I mean, using TCP/IP and PPP as transport level ? >Just about to next week! > >> >> What is the kind of speed you get ? >> Any experience with other baud rates over WAN, like 56..64 kbit/s ? >Yep. I have 12 SCO machines calling online on a 13th. I have 3 of these >machines at the far end of 1 64K link and a fourth at the end of another >64K piggy backed on the end of the first. A total of 90+ remote users. >Very good performance! It's a telephone data entry application in >the travel business. Best move I made going WAN. Previously all >the machines were in one location and the remote terminals were connected >using X25 pads over the same 64K lines. The performance was pretty bad. > >Clive Eisen Email ceisen@deepthwt.demon.co.uk >Clive Eisen Computing Phone (+44) 71 281 1414 >From: uunet!cssnet.sanford.nc.us!david (J. David Schronce) > >lerner@ecitele.com (Yaakov Lerner (SDH-S)) writes: > >> >> In practice, did anybody try or use SQL database access >> over 9.6 kbit/s line ? I mean, using TCP/IP and PPP as transport level ? >> >> What is the kind of speed you get ? >> Any experience with other baud rates over WAN, like 56..64 kbit/s ? >> > >I'm running 6 remote sites of 56K lines without any problem. 2 of them >telnet in over the Internet. Of course, I'm feeding character terminals. > I'm afraid X, even at 56K would be _WAY_ too slow! Perhaps some of the >new X products will remedy this problem, like the NCD drivers or the Qume >serial based X terminals... > >---dave--- > >david@cssnet.sanford.nc.us [J. David Schronce] >Computer Support Services - Public Access Usenet - 919/708-5066 Jacob Lerner ECI Telecom, Israel lerner@ecitele.com Tel. +972(3)926-6254