Distributed computing backup and recovery system

• US 8,930,320 B2
• Filed: 09/30/2011
• Issued: 01/06/2015
• Est. Priority Date: 09/30/2011
• Status: Active Grant

• Current Assignee: Accenture Global Services Limited (Accenture PLC)
• Original Assignee: Accenture Global Services Limited (Accenture PLC)
• Inventors: Tung, Teresa, Richter, Owen E., Farooqui, Sameer
• Primary Examiner(s): Le, Miranda
• Application Number:
  US13/250,901
• Publication Number:
  US 20130085999A1
• Time in Patent Office: 1,194 Days
• Field of Search: 707/654, 707/999.204, 707/698
• US Class Current: 707/654
• CPC Class Codes: G06F 11/1451 by selection of backup contents
• CPC Class Codes: G06F 11/1464 for networked environments
• CPC Class Codes: G06F 11/1469 Backup restoration techniques
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor

Abstract

The distributed computing backup and recovery (DCBR) system and method provide backup and recovery for distributed computing models (e.g., NoSQL). The DCBR system extends the protections from server node-level failure and introduces persistence in time so that the evolving data set may be stored and recovered to a past point in time. The DCBR system, instead of performing backup and recovery for an entire dataset, may be configured to apply to a subset of data. Instead of keeping or recovering snapshots of the entire dataset which requires the entire cluster, the DCBR system identifies the particular nodes and/or archive files where the dataset resides so that backup or recovery may be done with a much smaller number of nodes.

Claims

1. A method for distributed computing backup and recovery, comprising:

retrieving at least one user selectable preference;

identifying a first subset of data from within a data set according to the at least one user selectable preference, the first subset of data containing less than all of the data in the data set wherein the first subset is selectable by using the user selectable preference;

receiving, into a memory via an interface controlled by a processor connected to a network in a computing environment wherein the identified data objects are within a second subset of data, the second subset of data containing less than all of the data in the first subset of data;

evaluating, using the processor, a hash function stored in the memory to determine network storage locations or network retrieval locations, or both for the data objects;

storing at a granular level, at each of the network storage locations, the data objects according to the data object request, when the data object request comprises a request to store the data objects, where the stored data objects are identified as a replica of the data objects stored at each of the network storage locations;

retrieving a hash seed for an identified time used to recreate the hash function for the identified time,

where the data object request comprises the identified time to retrieve the data objects;

where the hash seed is a random function based on an initial seed, where the hash seed was previously stored for the identified time;

recreating, using the hash seed, the hash function for the identified time;

retrieving a backup record identifier from a backup log file corresponding to the data object request;

applying the hash function to the backup record identifier, where the hash function identifies the network retrieval locations in the computing environment the data objects are stored;

identifying the data objects within the second subset of data;

retrieving at a granular level from one of the network retrieval locations from a backup of the computing environment, using the processor connected to the network, the stored data objects identified by the one of the network retrieval locations, when the data object request comprises a request to retrieve the data objects, where the stored data objects are retrieved from the second subset of data;

retrieving one of the replica of the data objects from a node where the hash function determines where the stored data objects are located of the network retrieval locations; and

returning the data objects responsive to the request.

2. The method of claim 1, further comprising:

applying the hash function to determine, according to selection criteria, a preferred replica among the replica of the data objects retrieved from nodes in the computing environment.

3. The method of claim 2, where the selection criteria comprises data freshness, data correctness of the data objects, or a combination of both.

4. The method of claim 1, further comprising:

determining a hash seed used to recreate the hash function for an identified time and storing the hash seed for a identified time, at one or more of the network storage locations,

where the hash seed for the identified time is used to recreate the hash function for the identified time,

where the hash seed is a random function based on an initial seed; and

when the data object request comprises a request to store the data objects, recording write accesses that occur in the computing environment until the data objects are stored; and

when the data object request comprises a request to retrieve the data objects, playing back the recording of write accesses until a restore completes.

5. The method of claim 1, wherein retrieving further comprises determining a plurality of configurable restore options, the restore options including:

a sequence of a plurality of nodes to restore including one or more nodes from which to retrieve a copy of the data objects;

or

restore nodes to use to restore the data objects concurrently;

or

a combination thereof.

6. A product for distributed computing backup and recovery, comprising:

a computer readable memory with processor executable instructions stored thereon, wherein the instructions when executed by the processor cause the processor to;

retrieve at least one user selectable preference;

identify a first subset of data from within a data set according to the at least one user selectable preference, the first subset of data containing less than all of the data in the data set wherein the first subset is selectable by using the user selectable preference;

receive, into a memory via an interface controlled by a processor connected to a network in a computing environment, a data object request that identifies data objects to store or retrieve, wherein the identified data objects are within a second subset of data, the second subset of data containing less than all of the data in the first subset of data;

evaluate, using the processor, a hash function stored in the memory to determine network storage locations or network retrieval locations, or both for the data objects;

identify the data objects within the second subset of data;

store at a granular level, at each of the network storage locations, the data objects according to the data object request, when the data object request comprises a request to store the data objects, where the stored data objects are identified as a replica of the data objects at each of the network storage locations;

retrieve a hash seed for an identified time used to recreate the hash function for the identified time,

where the hash seed is a random function based on an initial seed, where the hash seed was previously stored for the identified time,

where the data object request comprises the identified time to retrieve the data objects;

recreate, using the hash seed, the hash function for the identified time;

retrieve a backup record identifier from a backup log file corresponding to the data object request;

apply the hash function to the backup record identifier, where the hash function identifies the network retrieval locations in the computing environment the data objects are stored;

retrieve at a granular level, from one of the network retrieval locations from a backup of the computing environment, using the processor connected to the network, the data objects identified by the one of the network retrieval locations, when the data object request comprises a request to recover or retrieve the data objects, where the stored data objects are retrieved from the second subset of data;

retrieve the replicas of the data objects from nodes where the hash function determines where the stored data objects are located of the network retrieval locations; and

return the data objects responsive to the request.

7. The product of claim 6, further comprising analyzing the replicas of the data objects retrieved from nodes in the computing environment to determine a preferred replica from the retrieved replicas.

8. The product of claim 7, wherein the preferred replica comprises selection criteria that includes:

data freshness, or data correctness of the data objects, or a combination of both.

9. The product of claim 6, the instructions when executed by the processor further cause the processor to:

determine a hash seed used to recreate the hash function for an identified time and store the hash seed for a identified time, at one or more of the network storage locations,

where the hash seed for the identified time is used to recreate the hash function for the identified time, where the hash seed is a random function based on an initial seed;

record write accesses that occur in the computing environment until the data objects are stored, when the data object request is a request to store the data objects; and

play back the recording of write accesses until a restore completes, when the data object request is a request to retrieve the data objects.

10. The product of claim 6, the instructions when executed by the processor further cause the processor to determine a plurality of configurable restore options, the restore options including:

a sequence of a plurality of nodes to restore including one or more nodes from which to retrieve a copy of the data objects;

or

restore nodes to use to restore the data objects concurrently;

or

a combination thereof.

11. A system for distributed computing backup and recovery (DCBR), comprising:

a processor to retrieve at least one user selectable preference, identify a first subset of data from within a data set according to the at least one user selectable preference, the first subset of data containing less than all of the data in the data set wherein the first subset is selectable by using the user selectable preference;

a cluster of computing nodes in a computing environment;

an interface controlled by the processor connected to a network in the computing environment;

a memory coupled to the processor, wherein the memory comprises;

a data object request received through the interface for data objects wherein the data objects are within a second subset of data, the second subset of data containing less than all of the data in the first subset of data;

a hash function that is evaluated by the processor to determine network storage locations or network retrieval locations, or both for the data object;

instructions executable by the processor that cause the processor to;

retrieve a hash seed for an identified time used to recreate the hash function for the identified time,

where the hash seed is a random function based on an initial seed, where the hash seed was previously stored for the identified time,

where the data object request comprises the identified time to retrieve the data objects;

recreate, using the hash seed, the hash function for the identified time;

retrieve a backup record identifier from a backup log file corresponding to the data object request;

apply the hash function to the backup record identifier, where the hash function identifies the network retrieval locations in the computing environment the data objects are stored;

identify the data objects within the second subset of data;

retrieve at a granular level from one of the network retrieval locations the data objects from a backup of the computing environment, when the request is a request to retrieve the data objects, where the stored data objects retrieved are identified by the one of the network retrieval locations, where the stored data objects are retrieved from the second subset of data;

store at a granular level the data objects, when the request is a request to store the data object;

where a copy of the data objects are located on one or more of the nodes, where the stored data objects are identified as a replica of the data objects at each of the network storage locations;

retrieve the replicas of the data objects from nodes where the hash function determines where the stored data objects are located of the network retrieval locations; and

return the data objects responsive to the request.

12. The system of claim 11, where the instructions further cause the processor to retrieve one of the replica of the data objects from a node where the hash function determines where the stored data objects are located of the network retrieval locations.

13. The system of claim 12, further comprises a preferred replica determined according to selection criteria that includes:

data freshness, or data correctness of the data objects, or a combination of both.

14. The system of claim 13, wherein the instructions further cause the processor to determine the preferred replica by applying the hash function to the retrieved one of the replica.

15. The system of claim 11, further comprises:

a backup log file that includes a backup record identifier corresponding to the data object request that the hash function uses to identify where in the computing environment the data objects are stored.

16. The system of claim 11, further comprises:

a preferred replica from the replicas, the preferred replica comprising the most recent of the data objects, or the most correct copy of the data objects, or a combination of both.

17. The system of claim 11, where the instructions further cause the processor to execute restore options, the restore options including:

a sequence of a plurality of nodes to restore including the one or more nodes from which to retrieve the copy of the data objects;

or

restore nodes to use to restore the data objects concurrently;

or a combination thereof.