Method of transferring messages between computer programs across a network

• US 6,817,018 B1
• Filed: 05/21/1997
• Issued: 11/09/2004
• Est. Priority Date: 10/08/1993
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Johnson, Peter, Clarke, Paul, Kingston, William, Black, George, Drew, Robin Miles, Meli, Roger
• Primary Examiner(s): LAO, SUE X
• Application Number:
  US08/861,181
• Time in Patent Office: 2,729 Days
• Field of Search: 395/671, 395/182.14, 395/200.62, 709/101
• US Class Current: 719/313
• CPC Class Codes: G06F 11/1443 Transmit or communication errors
• CPC Class Codes: G06F 2201/82 Solving problems relating to consistency
• CPC Class Codes: G06F 9/466 Transaction processing
• CPC Class Codes: G06F 9/546 Message passing systems or structures, e.g. queues

Abstract

A method of delivering messages between application programs is provided which ensures that no messages are lost and none are delivered more than once. The method uses asynchronous message queuing. One or more queue manager programs (100) is located at each computer of a network for controlling the transmission of messages to and from that computer. Messages to be transmitted to a different queue manager are put onto special transmission queues (120). Transmission to an adjacent queue manager comprises a sending process (130) on the local queue manager (100) getting messages from a transmission queue and sending them as a batch of messages within a syncpoint-manager-controlled unit of work. A receiving process (150) on the receiving queue manager receives the messages and puts them within a second syncpoint-manager-controlled unit of work to queues (180) that are under the control of the receiving queue manager. Commitment of the batch is coordinated by the sender transmitting a request for commitment and for confirmation of commitment with the last message of the batch, commit at the sender then being triggered by the confirmation that is sent by the receiver in response to the request.

The invention avoids the additional message flow that is a feature of two-phase commit procedures, avoiding the need for resource managers to synchronise with each other. It further reduces the commit flows by permitting batching of a number of messages.

Claims

1. A method of transactional control of message transfer across a transaction-oriented data processing network wherein a sender program is responsible for sending messages from a first node of the network and a receiver program is responsible for receiving messages at a second node of the network, the method comprising:

sending messages by the sender program within a first syncpoint-manager-controlled unit of work and receiving messages by the receiver program within a second syncpoint-manager-controlled unit of work, while holding the sending and receiving operations in-doubt, uncommitted, until resolution of the first and second units of work, respectively, wherein the first syncpoint-manager-controlled unit of work and the second syncpoint-manager-controlled unit of work are logically linked so that commit processing at resolution of said units of work comprises the steps of;

in response to successful receipt of the messages by the receiver program, performing the sequence of steps of committing said second unit of work, transmitting to the sender program a positive confirmation of receipt, and in response to the positive confirmation committing the first unit of work;

or

in response to unsuccessful receipt of the messages, performing the sequence of steps of rolling back the second unit of work, transmitting to the sender program a negative confirmation of receipt, and in response to said negative confirmation backing out the first unit of work.

2. A method according to claim 1 wherein said sender and receiver programs are located on adjacent nodes within a network, and wherein messages, which may be destined for different destination nodes, are transmitted between adjacent nodes on the way to their respective destination nodes as a batch of messages within a unit of work, the units of work incorporating said sending and receiving operations being held in-doubt until the end of the batch.

3. A method according to claim 2 wherein the last message in a batch is transmitted together with a request for commitment of and for confirmation of receipt of the batch, the commitment of said second unit of work and the transmission of said positive or negative confirmation being in response to said request.

4. A method according to claim 1 wherein log records are written to record the in-doubt status of said units of work for use in recovery processing following a failure during the processing of said units of work, the log records being read during recovery processing to determine which units of work should be committed and which should be backed out.

5. A method according to claim 1, which implements messaging and queuing for communication between application programs, the application programs sending messages to message queues from where receiver application programs can asynchronously take the messages for processing or forwarding on.

6. A method according to claim 5, wherein communication between application programs running on different computer systems of the network comprises at least the following steps:

a first application program issuing a put message instruction under control of a synchpoint manager in the sending computer system, for sending a message to a message queue;

sender and receiver transmission programs transferring messages between the computer systems, as two logically linked units of work, using synchpoint managers in both the sending and receiving computer systems; and

a second application program issuing a get message instruction under control of a synchpoint manager in the receiving computer system, for taking the message from the queue;

wherein the operations of put message, transfer and get message are each held in-doubt until resolution of the respective unit of work.

7. A data processing system including a messaging manager for transactional control of message transfer across a network of data processing systems, the messaging manager including sender and receiver programs for transferring messages between adjacent messaging managers in the network in accordance with the following transfer protocol:

a sender program of a first messaging manager sending one or more messages within a first syncpoint-manager-controlled unit of work;

a receiver program in a second messaging manager receiving said messages within a second syncpoint-manager-controlled unit of work;

the sending and receiving operations being held in-doubt, uncommitted, until resolution of the first syncpoint-manager-controlled unit of work and the second syncpoint-manager-controlled unit of work, respectively; and

said first and second units of work being logically linked so that commit processing at resolution of said first and second units of work comprises either

(i) in response to successful receipt of the messages by the receiver program, committing said second unit of work, transmitting to the sender program a positive confirmation of receipt, and in response to the positive confirmation committing the first unit of work;

or

(ii) in response to unsuccessful receipt of the messages, rolling back the second unit of work, transmitting to the sender program a negative confirmation of receipt, and in response to said negative confirmation backing out the first unit of work.

8. A data processing system according to claim 7, wherein the messaging manager is adapted for message queuing inter-program communication across a heterogeneous network of data processing systems, the messaging manager including an application programming interface by which applications attach to the messaging manager and providing queuing services enabling application programs to put messages onto message queues for asynchronous retrieval by other application programs.

9. A computer program product comprising computer program code recorded on a computer-readable recording medium, the computer program code implementing a messaging manager for transactional control of message transfer across a network of data processing systems, the messaging manager including sender and receiver programs for transferring messages between adjacent messaging managers in the network in accordance with the following transfer protocol:

a sender program of a first messaging manager sending one or more messages within a first syncpoint-manager-controlled unit of work;

a receiver program in a second messaging manager receiving said messages within a second syncpoint-manager-controlled unit of work;

the sending and receiving operations being held in-doubt, uncommitted, until resolution of the first syncpoint-manager-controlled unit of work and the second syncpoint-manager-controlled unit of work, respectively; and

said first and second units of work being logically linked so that commit processing at resolution of said first and second units of work comprises either

(i) in response to successful receipt of the messages by the receiver program, committing said second unit of work, transmitting to the sender program a positive confirmation of receipt, and in response to the positive confirmation committing the first unit of work;

or

(ii) in response to unsuccessful receipt of the messages, rolling back the second unit of work, transmitting to the sender program a negative confirmation of receipt, and in response to said negative confirmation backing out the first unit of work.