Home » » Connection Establishment and Clearing

Connection Establishment and Clearing

Written By nuriyati mauji on Monday, 21 October 2013 | Monday, October 21, 2013

Connection Establishment and Clearing

  To identify the separate data streams that a TCP may handle, the TCP  provides a port identifier.  Since port identifiers are selected  independently by each TCP they might not be unique.  To provide for  unique addresses within each TCP, we concatenate an internet address  identifying the TCP with a port identifier to create a socket which  will be unique throughout all networks connected together.
  A connection is fully specified by the pair of sockets at the ends.  A  local socket may participate in many connections to different foreign  sockets.  A connection can be used to carry data in both directions,  that is, it is "full duplex".
  TCPs are free to associate ports with processes however they choose.  However, several basic concepts are necessary in any implementation.  There must be well-known sockets which the TCP associates only with  the "appropriate" processes by some means.  We envision that processes  may "own" ports, and that processes can initiate connections only on  the ports they own.  (Means for implementing ownership is a local  issue, but we envision a Request Port user command, or a method of  uniquely allocating a group of ports to a given process, e.g., by  associating the high order bits of a port name with a given process.)


Post a Comment