In a ring network where network nodes are interconnected by an optical fiber ring, one of the nodes acts as a master node and each of the other nodes acts as a slave node. The master node successively designates a different slave node in a descending order of priorities, transmits a command message to the designated slave node, and updates its memory with a response message from the designated node. If the command message is received, each designated slave node updates its memory with data contained in the command message and returns a response message to the master node. If the command message is not received, each slave node, acting as a temporary master node, designates a different one of the other slave nodes in descending order, transmits a command message to each designated node, and updates its memory with a response message therefrom. Each temporary master node repeats the process until it receives a command message from another temporary master node having a priority higher than the priority of the sending node.

A session recovery mechanism that permits the recovery of a session with a minimal delay to a user and with minimal data loss. When the client/server communications protocol process, such as TCP/IP process, issues an error message to a server and a client, the server and the client switch from a server data socket and a client data socket, respectively, to a new server data socket and a new client data socket, respectively. This switchover is achieved by having the client open a listening socket during its initialization process. Using the client listening socket, the client listens for a connection from the server to switch to a new data socket, in case of, for example, error messages from the TCP/IP process.

A method and apparatus of releasing resources allocated to terminated sessions at a secondary data processing node. A first node transmits one or more switch messages to a second node requesting the second node to redistribute sessions on a first route between the nodes among one or more alternate routes. Each switch message identifies the sessions to be switched and an alternate route to which the sessions are to be switched. The last switch message transmitted by the first node is indicated to be the last message in some appropriate way and includes a count of the total number of sessions that have been requested to be switched. The second node accumulates a count of the total number of sessions on the first route for which it has received a switch request. The second node compares its accumulated count with the total count derived from the last switch message and releases the resources allocated to all remaining sessions on the first route in the event the counts are equal.

In a distributed network system, resilient virtual fault tolerant service sessions are conducted between a host application program and a telnet client user over an end-to-end link including a telnet server, a messaging protocol process, a network driver, and a communication link. The telnet server establishes a tty type structure in response to a service request from a telnet user/client. This invention includes a service control block which controls the establishment of a tty and associated request control blocks and buffers for processing I/O requests. The session is functionally partitioned into an upper half session involving the host application program and the telnet server, and a lower half session involving the remainder of the network components leading to the telnet client. When an interruption in a session occurs due to a fault originating in those portions of the network devoted to the lower half session, the telnet server continues to operate the upper half session by accepting I/O requests from the host application program and queueing these requests in the request control buffers, up to a maximum number of permitted queued requests. The host file system determines the handling of requests that exceed the maximum.

A method of operation of a data communications system (5) with session control having a data communications network (10) and at least a first communications device (1), a second communications device (2) and a third commnunications device (3) in communications with the data communications network (10). The data communications network (10) identifies a fault status change in communications with the first communications device (1), generates a session control command (6) with an identification number (7) attached responsive to identifying the fault status change and sending the session control command (6) to at least the second and the third communications devices (2 and 3). At least at the second communications device (2), the second communications device (2) receives the session control command (6), deter-mines from the identification number (7) whether the second communications device (2) is affected by the fault status change and selectively performs a session control action in response thereto.