Browse Prior Art Database

Automatic Assignment of Unique Client Identifiers

IP.com Disclosure Number: IPCOM000122978D
Original Publication Date: 1998-Mar-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 2 page(s) / 69K

Publishing Venue

IBM

Related People

Shields, I: AUTHOR [+2]

Abstract

This problem addresses the problem of servers maintaining a persistent unique identifier for a client computer within a TCP/IP network where it is assumed that the client IP address can be assigned dynamically each time the client computer connects to the network. Servers must have a persistent identifier of their client computers if the server is required to maintain client/server context (i.e., state information) from one network connection to the next.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Automatic Assignment of Unique Client Identifiers

      This problem addresses the problem of servers maintaining a
persistent unique identifier for a client computer within a TCP/IP
network where it is assumed that the client IP address can be
assigned dynamically each time the client computer connects to the
network.  Servers must have a persistent identifier of their client
computers if the server is required to maintain client/server context
(i.e., state information) from one network connection to the next.

      The approach is for the client (C) and the server (S) to
maintain a persistent (e.g., disk) copy of a client identifier that
is automatically generated by S.

      The protocol works as follows:

      Assume that C initiates a connection with S. First, C looks to
see if it has a persistent identifier (pI).  If it does, it sends pI
to S.  Upon receiving pI, S looks to see if it has recorded (on
persistent storage) pI as one of its client identifiers.  If the
server finds pI, it  sends an indication to the client that pI is
okay.  If the server cannot  find pI, it automatically generates a
new identifier (pI') that guarantees uniqueness among all that
clients that have connected to the  server.  One unique identifier
generation technique is to use a time-stamp concatenated with a
sequence number.  The server, S, returns  pI' to C, indicating  that
pI was not valid at S.  If the client, C, receives a new identifier,
it must clean-up any resources the that had  been previously
allocated during previous connections with the server,  S; for
example, the old identifier, pI, should be removed from persistent
storage.

      The first time C attempts to connect to S, it will not have an
identifier, pI.  In this case, C sends a...