Method for determining a recovery schedule

• US 20080126279A1
• Filed: 08/22/2006
• Published: 05/29/2008
• Est. Priority Date: 08/22/2006
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Santos, Cipriano, Brau, Ernesto, Keeton, Kimberly, Merchant, Arif, Zhang, Alex, Beyer, Dirk
• Granted Patent: US 7,426,616 B2
• Time in Patent Office: Days
• US Class Current: 706/19
• CPC Class Codes: G06F 11/1469 Backup restoration techniques
• CPC Class Codes: G06F 11/2069 Management of state, configuration or failover
• CPC Class Codes: G06F 11/2082 Data synchronisation

Abstract

Provided is a method for determining a recovery schedule. The method includes accepting as input a recovery graph. The recovery graph presents one or more strategies for data recovery. In addition, at least one objective is provided and accepted. The recovery graph is formalized as an optimization problem for the provided objective. When formalized as an optimization problem, at least one solution technique is applied to determine at least one recovery schedule.

Claims

1. A method for determining a recovery schedule, comprising:

accepting as input a recovery graph, the recovery graph presenting one or more recovery strategies for data recovery;

accepting at least one objective for the recovery schedule;

formalizing the recovery graph as an optimization problem for the objective; and

applying at least one solution technique to the optimization problem to determine at least one recovery schedule.

2. The method of recovering data of claim 1, wherein each recovery strategy is a path within the recovery graph, each path including a series of nodes, each node indicating a recovery task or operational state, each node further indicating at least one resource demanded by the task or state, and precedence constraints represented by edges between the nodes.

3. The method of recovering data of claim 1, wherein the applied solution technique is selected from the group consisting of:

mixed integer programming with an off-the-shelf solver, constraint programming with an off-the-shelf solver, randomized algorithms, genetic algorithms, domain-specific heuristics, and combinations thereof.

4. The method of recovering data of claim 1, wherein the objective of the optimization problem is selected from the group consisting of:

minimizing financial penalties, minimizing recovery time, minimizing data loss, minimizing vulnerability to subsequent failure, minimizing disruption to unaffected workloads, satisfying recovery time objectives, satisfying data loss objectives, satisfying scheduling deadlines and combinations thereof.

5. The method of recovering data of claim 1, wherein an optimized solution decomposes the problem into two parts, a first part including path selection for each workload, and a second part including schedule determination for each selected path, the first part and the second part each having an applied solution technique.

6. The method of recovering data of claim 1, wherein the method is stored on a computer-readable medium as a computer program that, when executed by a computer, will perform the steps of determining at least one recovery schedule.

7. A method for determining a recovery schedule for recovering datasets from dependable storage, comprising:

providing a recovery graph, the recovery graph presenting at least one recovery strategy for recovering application datasets, each recovery strategy represented as a path, each path including nodes, each node indicating a recovery task or operational state, each node further indicating at least one resource demanded by the task or state, precedence constraints represented by edges between the nodes;

accepting at least one objective for the recovery schedule;

formalizing the recovery graph into an optimization problem for the objective, the problem further receiving inputs, including penalty rates for each application dataset, and at least one capability for each resource; and

applying at least one solution technique to the optimization problem to determine at least one recovery schedule.

8. The method of recovering data of claim 7, wherein the penalty rates are selected from:

outage time penalties, recent data loss penalties, vulnerability penalties, disruption penalties, and combinations thereof.

9. The method of recovering data of claim 7, wherein the recovery graph includes paths for failed as well as non-failed application datasets.

10. The method of recovering data of claim 7, wherein the objective of the optimization problem is selected from the group consisting of:

minimizing financial penalties, minimizing recovery time, minimizing data loss, minimizing vulnerability to subsequent failure, minimizing disruption to unaffected workloads, satisfying recovery time objectives, satisfying data loss objectives, satisfying scheduling deadlines and combinations thereof.

11. The method of recovering data of claim 7, wherein an optimized solution decomposes the problem into two parts, path selection and schedule determination, each having an applied solution technique.

12. The method of recovering data of claim 7, wherein providing the recovery graph includes:

determining application workloads in a system environment;

determining resources within the environment and their capabilities;

determining data protection and recovery techniques available for each application;

determining steps for each recovery strategy, including for each step the resource demands and ordering constraints; and

assembling the recovery graph from the determined resources, application workloads, data protection and recovery techniques, recovery strategy steps and a disaster scenario.

13. The method of recovering data of claim 7, wherein formalizing the recovery graph into an problem includes:

determining precedence constraints for the recovery of each application dataset and reinstatement of data protection workloads for each application dataset;

determining resource constraints, where the sum of the demands for a resource under a candidate schedule can not exceed the capabilities of the resource.

14. The method of recovering data of claim 7, wherein the application dataset includes application binaries, application metadata and user data.

15. A method for determining a recovery schedule for recovering data, comprising:

accepting as input a recovery graph, the recovery graph presenting one or more recovery strategies for recovering application datasets, each recovery strategy represented as a path, each recovery action for each path represented as a node along each path, precedence constraints represented by directed edges between the nodes, and resource demands represented as node annotations;

obtaining the penalty rates associated with each application dataset;

obtaining resource demands for each recovery action for each recovery strategy for each application dataset to be recovered;

obtaining the capabilities of resources within the environment;

obtaining at least one objective for the recovery schedule; and

applying at least one solution technique incorporating the recovery graph, the penalty rates and the resource demands to determine at least one recovery schedule for the objective.

16. The method of recovering data of claim 15, wherein determining the resource demands for each recovery action includes the resource demands for each failed and non-failed workload.

17. The method of recovering data of claim 15, wherein in applying the solution technique the recovery graph is formalized as an optimization problem for the objective.

18. The method of recovering data of claim 15, wherein the penalty rates are based on one or more factors selected from the group consisting of:

choice of path, start time of a job, end time of a job, duration of a job, involved resources, and combinations thereof.

19. The method of recovering data of claim 15, wherein an optimized solution decomposes the problem into two parts, a first part including path selection for each workload, and a second part including schedule determination for each selected path, the first part and the second part each having an applied solution technique.

20. The method of recovering data of claim 15, wherein the method is stored on a computer-readable medium as a computer program which, when executed by a computer will perform the steps of determining at least one recovery schedule.