# US 2002/0178395 A1 · Multi-agent cooperative transaction method and system

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2002/0178395 A1
- Document ID: US20020178395A1
- Status: Published application
- Kind: A1
- Application number: 09/863,486
- Filed: 2001-05-23
- Published: 2002-11-28
- Earliest claimed: 2001-05-23
- Assignees: Hewlett-Packard Development Company, L.P.
- Inventors: Qiming Chen; Umeshwar Dayal
- Web page: https://patentgrep.com/patent/US20020178395A1

## CPC

- H04M 15/74: {Backing up}
- H04M 15/00: Arrangements for metering, time-control or time indication {; Metering, charging or billing arrangements for voice wireline or wireless communications, e.g. VoIP}
- H04M 2215/709: Backup
- H04M: TELEPHONIC COMMUNICATION (circuits for controlling other apparatus via a telephone cable and not involving telephone switching apparatus G08)

## Abstract

A method and system for processing multi-agent cooperative transactions. A failure detector is provided for detecting whether a failure is an intra-enterprise failure or an inter-enterprise failure. An intra-enterprise failure handler is coupled to the failure detector for performing failure recovery for intra-enterprise failures. Failure recovery for intra-enterprise failures can include identifying the scope of failure recovery within a first enterprise. Once the scope of failure recovery has been identified, a top-down undo operation of sub-transactions in the identified scope may be performed within the first enterprise. An inter-enterprise failure handler is also coupled to the failure detector for performing failure recovery for inter-enterprise failures. Failure recovery for inter-enterprise failures can include identifying the scope of failure recovery in a second enterprise to which a failure in a first enterprise has been transfered. Once the scope of failure recovery has been identified, a top-down undo operation of sub-transactions in the identified scope may be performed in the second enterprise.

## Claims

### Claim 1

A system for processing multi-agent cooperative transactions comprising:
a) a failure detector for detecting whether a failure is an inter-enterprise failure or an intra-enterprise failure;
b) an intra-enterprise failure handler coupled to the failure detector for performing failure recovery for intra-enterprise failures; and
c) an inter-enterprise failure handler coupled to the failure detector for performing failure recovery for inter-enterprise failures.

### Claim 2 (depends on 1)

The system of claim 1 wherein the intra-enterprise failure handler includes
a scope determination module for identifying the failure recovery scope; and
a top-down logical undo module coupled to the scope determination module for undoing sub-transactions in the identified scope in a top-down manner.

### Claim 3 (depends on 2)

The system of claim 2 wherein the scope determination module terminates at a closest ancestor of the failed transaction that is one of non-vital and associated with a contingency transaction.

### Claim 4 (depends on 1)

The system of claim 1 wherein the scope determination module when the failed transaction is non-vital to its parent, continuing with the parent transaction;
when is determined to be a result of agent malfunction, the parent transaction delegating the sub-transaction to another agent; and
when the failed transaction is associated with a contingency transaction, retrying the contingency transaction.

### Claim 5 (depends on 1)

The system of claim 1 wherein the inter-enterprise failure handler includes a scope determination module for identifying the failure recovery scope; and
a top-down logical undo module coupled to the scope determination module for undoing sub-transactions in the identified scope in a top-down manner.

### Claim 6 (depends on 5)

The system of claim 5 wherein the top-down logical undo module when the transferred in transaction has not started, then not performing failure handling on the transferred in transaction;
when the transferred in transaction is in progress, then determining the rollback root of the transferred in transaction; and
when the transferred in transaction has completed, then compensating for the transferred in transaction.

### Claim 7 (depends on 1)

The system of claim 1 wherein the top-down logical undo module when the rollback root of the transferred in transaction, then utilizing the rollback root as the root of recovery;
when the rollback root of the transferred in transaction has committed to one of a parent and an ancestor, the parent is in-progress, and the rollback root of the transferred in transaction is one of non-vital and associated with a contingency transaction, then utilizing the parent of the rollback root as the root of recovery; and
when the rollback root of the transferred in transaction and the parent of the rollback root have committed, then determining the highest committed ancestor of the rollback root of the transferred in transaction and determining whether the rollback root of the highest committed ancestor is in progress;
when the highest committed ancestor is in progress, the highest committed ancestor of the rollback root of the transferred in transaction is utilized as an extended rollback root;
when the highest committed ancestor is not in progress, the parent of the highest committed ancestor of the rollback root of the transferred in transaction is utilized as an extended rollback root.

### Claim 8

A method of processing multi-agent cooperative transactions comprising the steps of:
a) a first transaction waiting for a commit notification; the first transaction including a dependency on a second transaction;
b) the second transaction sending a commit notification; and
c) the first transaction receiving the commit notification and committing only after receipt of the commit notification.

### Claim 9 (depends on 8)

The method of claim 8 wherein the dependency is a start dependency; and
wherein the first transaction does not initiate processing until the second transaction is initiated.

### Claim 10 (depends on 8)

The method of claim 8 wherein the dependency is a failure dependency;
wherein the first transaction aborts when the second transaction aborts.

### Claim 11 (depends on 8)

The method of claim 8 wherein the dependency between the first transaction and a second transaction is a mutual settle dependency; and wherein the first transaction and the second transaction can commit only when the first transaction and the second transaction have reached agreement on a contract.

### Claim 12 (depends on 8)

The method of claim 8 wherein the dependency between the first transaction and a second transaction is a mutual abort-compensate dependency; and
wherein when a first transaction fails, then the second transaction is either aborted or compensated for.

### Claim 13 (depends on 8)

The method of claim 8 wherein sending a commit notification employs a point-to-point communication.

### Claim 14 (depends on 8)

The method of claim 8 wherein sending a commit notification employs a multicast communication.

### Claim 15 (depends on 8)

The method of claim 8 wherein the step of the first transaction receiving the commit notification and committing only after receipt of the commit notification includes
committing to a first database associated with the first transaction.

### Claim 16 (depends on 8)

The method of claim 8 wherein a centralized coordinator for controlling the cooperation between transactions is not needed.

### Claim 17

A method for processing failures in multi-agent cooperative transactions comprising:
a) detecting whether a failure is an inter-enterprise failure or an intra-enterprise failure;
b) when the failure is an intra-enterprise failure, performing failure recovery for the intra-enterprise failure; and
c) when the failure is an inter-enterprise failure, performing failure recovery for the inter-enterprise failure.

### Claim 18 (depends on 17)

The method of claim 17 wherein the step of when the failure is an intra-enterprise failure, performing failure recovery for the intra-enterprise failure includes
identifying the failure recovery scope; and
undoing sub-transactions in the identified scope in a top-down manner.

### Claim 19 (depends on 17)

The method of claim 17 wherein the step of when the failure is an inter-enterprise failure, performing failure recovery for the inter-enterprise failure includes
identifying the failure recovery scope; and
undoing sub-transactions in the identified scope in a top-down manner.

### Claim 20 (depends on 19)

The method of claim 19 wherein identifying the failure recovery scope includes the step of terminating at a closest ancestor of the failed transaction that is one of non-vital and associated with a contingency transaction.
