[84] in resnet

home help back first fref pref prev next nref lref last post

[daemon@ATHENA.MIT.EDU : the rx hex]

daemon@ATHENA.MIT.EDU (marc@MIT.EDU)
Thu Nov 11 00:39:09 1993

From: marc@MIT.EDU
Date: Thu, 11 Nov 93 00:33:48 EST
To: resnet@MIT.EDU


------- Forwarded transaction

[2178]  daemon@ATHENA.MIT.EDU (peter honeyman) Info-AFS_Redistribution 11/10/93 22:18 (41 lines)
Subject: the rx hex
From: peter honeyman <honey@citi.umich.edu>
To: info-afs@transarc.com
Date: Wed, 10 Nov 1993 18:29:56 -0500

here's the long-awaited (long delayed, actually) tech report describing
some of our rx dialup hacks at citi.  postscript is in

  /afs/citi.umich.edu/user/honey/papers/rxp/rx.tr.PS

(it will be in "the usual place" soon.)

	peter

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=


                       CITI Technical Report 93-8

                               The Rx Hex

                              D. Bachmann
                        bachmann@austin.ibm.com

                              P. Honeyman
                          honey@citi.umich.edu

                              L.B. Huston
                         lhuston@citi.umich.edu

                                ABSTRACT

      At CITI, we run dataless AFS clients  over  dialup  IP  net-
      works.   Improving  Rx performance is critical to that task.
      In this paper, we report on our progress in adapting  Rx  to
      networks  characterized  by  low  bandwidth,  high delay, or
      variable round-trip time.  Our focus is on adding facilities
      for  congestion avoidance and control, and on compressing Rx
      headers.  Although our work is  ongoing,  we  have  overcome
      several   hurdles,  and  are  now  getting  impressive  data
      transfer rates over ordinary dialup lines.

--[2178]--

------- End forwarded transaction

home help back first fref pref prev next nref lref last post