NETWORK WORKING GROUP                                   R. L. Sunberg
Request for Comments #133                               Harvard University
IC 6710                                                 27 April 1971

                   FILE TRANSFER AND ERROR RECOVERY

   [Categories C.4, C.5, C.6, D.4, D.7, D.7]
1   FILE TRANSFER PROTOCOL

1A  Handshaking

1A1 I think that Mr Bhushan(RFC #114, NIC 5823) is not strict enough in

   his concept of a transaction sequence.  Every transaction should

   prompt a response from its recipient )recall Kalin's crates --

   RFC #60, NIC 4762).  Control should pass back and forth until the

   server terminates.  The server  always gets the last word (more on

   error recovery later).

1A2 Some sample interchanges are given.

   User                Server          Comments
   <...>       ==>                     Establish a connection
               <==     <...>
   <I><...>    ==>                     Identify self
               <==     <+>             Ok, ready

   <R><...>    ==>                     Retrieval request
               <==     <rs>            I've got your file
   <rr>        ==>                     Send it
               <==     <,><...>        Here's the first part
   <rr>        ==>                     Got it
               <==     <+>             All done

   <S><...>    ==>                     Store request
               <==     <rr>            Ok, go ahead
   <#><...>    ==>                     Here's some protection stuff
               <==     <rr>            Ok
   <*><...>    ==>                     Here's the file
               <==     <+>             Got it.  All done.







                                                               [Page 1]

NWG/RFC #133


   See section 2B, below, for examples of error recovery.

1B  Extensions to the file transfer protocol

1B1 The file transfer protocol needs a mechanism for accessing

   individual records of a file.  This will be particularly useful

   when very large data bases appear on the network.  The following

   definitions should be added to the protocol:


   The store(s) and retrieve(R) requests have the data field format

   <key>, where <key> has the syntax:

   <key>::=<devicename>RS<filename>US<keyname>|<filename>US<keyname>.

     The <pathname> syntax is changed to:

   <pathname>::=<devicename>/<filename>/<pathname>RS<filename>.


   If a retrieve(R) request is given with a data field with <key>

   syntax rather than <pathname> syntax, then the returned data will

   consist of the record following the matching <key>.  If a store(s)

   request is given with a data field of <key> syntax, then the

   supplied data will replace the record following the matching

   <keyname>.  If the keyname does not exist, the record will be

   appended to the named file.  The individual installation must

   provide the linkage between the <keyname> and the record it

   references.


   In addition, the lookup(L) request will provide a list of keynames

   into a file (or the name of a file which contains the keynames).





                                                               [Page 2]

NWG/RFC #133


1B2 Transaction code F (request File directory) requests a listing of

   available files.  The data field of the F transaction is of the

   form:  <pathname>GS<pathname>GS...  All files in the server system

   which match one or move of the given <pathname> specifiers are

   listed in a return file.  The format of the data fiels of this

   file is:  <pathname>GS<pathname>GS...  If a <pathname> field in

   the request transaction does not include a <name> field, the

   default is all files on the given device.  Some examples are given:

   <F><DC1 DSK[62,50]] GS JOE>

   This example requests a list of all files on the disk specified by

   [62,50] plus all files named JOE.  The response could contain in

   the data field:

   <DC1 DSK[62,50] RS ALPHA RS BETA RS JOE GS DC1DSK[10,50] RS JOE>

   This message states that in the [62,50] area of the disk there are

   files ALPHA, BETA, and JOE, and that JOE is also a file in the

   [10,50] area of the disk.


2   ERROR RECOVERY


2A  Error recovery procedures have been noticeably lacking to date.

   The usual approach has been to close the connection and start from

   scratch.  Mr Bhushan proposes a third level abort but doesn't

   really detail the implementation.  I propose a multilevel error

   recovery procedure as follows.


2B  If an error occurs which does not cause a loss of third level



                                                               [Page 3]

NWG/RFC #133


   transaction boundaries and only affects one side of a duplex

   connection, a third level recovery is possible via a transaction

   sequence abort.  An example is given:

   User                Server          Comments
   <R><...>    ==>                     Send me this file
               <==     <rs>            Ok, I've got it
   <rr>        ==>                     Ready
               <==     <*><...error>   Here it is (with an error)
   <-><D>      ==>                     No.  (data) error
               <==     <-><D>          Sorry, forget it
   <R><...>    ==>                     Send the file (again)
               |<==    <rs>            Ready (doesn't get there)
               ...                     (waiting)
   <-><0>      ==>                     Error, timeout
               <==     <-><0>          Sorry, forget it
   <R><...>    ==>                     Send the file (third time)
               <==     <rs>            Got it
   <rr>        ==>                     Ready
               <==     <*><...>        There it is
   <rr>        ==>                     Got it
               <==     <+>             Done (finally>

   Note that the server always gets the last word in error situations

   as well as normal transmission.


2C  Although the above examples are given in terms of Bhushan's

   transaction codes, this form of error recovery is implementable in

   any protocol which uses flagged blocking and duplex connections.


2D  If errors cannot be recovered as above, then some meanst must be

   available to clear the link completely and resynchronize.  I

   suggest that an 8-bit argument be appended to the interrupt-on-link

   NCP message (INR, INS).  The receiver would send <INR><error> to

   indicate that the block boundaries were lost and all incoming data

   is being discarded.  The sender, upon receiving the INR, would



                                                               [Page 4]

NWG/RFC #133


   flush all queued output and wait for the link to clear.  The NCP

   would then send a <INS><newsync> message and, when it was received

   (RFNM returned), a negative termination would be sent on the link.

   The receiver begins accepting data again when the INS is received.

   This assumed that any process can flush untransmitted data and

   detect a clear link.  Note that this method is useable on any

   simplex connection.


2E  If all else fails, one can resort to closing the faulty socket.


3   NCP VERSION NUMBERS

3A  I suggest that the NCP be given a version number and the next

   version include two new message types:

       <WRU> ('Who aRe yoU?')  requests a version number from the

   receiving host and  <IAM><version> ('I AM')     supplies that

   number.


3B  The messages would probably be initially used in a 'can I talk to

   you?' sense or not at all.  Eventually, it would take on a 'what

   can you do?' meaning.  Accordingly, the <version> field should be

   large (32 bits?) for expansion.



      [ This RFC was put into machine readable form for entry ]
         [ into the online RFC archives by Jose Tamayo 4/97 ]








                                                               [Page 5]


You are viewing proxied material from ftp.icm.edu.pl. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.