Higher efficiency storage replication using compression

• US 9,560,093 B2
• Filed: 07/03/2014
• Issued: 01/31/2017
• Est. Priority Date: 03/08/2011
• Status: Expired due to Fees

• Current Assignee: Rackspace US Incorporated (Apollo Global Management, Inc.)
• Original Assignee: Rackspace US Incorporated (Apollo Global Management, Inc.)
• Inventors: Holt, Gregory Lee, Gerrard, Clay, Goetz, David Patrick, Barton, Michael
• Primary Examiner(s): Wong, Leslie
• Application Number:
  US14/323,726
• Publication Number:
  US 20150012491A1
• Time in Patent Office: 943 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1004 to protect a block of data words, e.g. CRC or checksum G06F11/1076 takes precedence; security arrangements for protecting computers or computer systems against unauthorized activity G06F21/00
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups
• CPC Class Codes: G06F 16/1744 using compression, e.g. sparse files
• CPC Class Codes: G06F 16/1748 De-duplication implemented within the file system, e.g. based on file segments de-duplication techniques in storage systems for the management of data blocks G06F3/0641
• CPC Class Codes: G06F 16/178 Techniques for file synchronisation in file systems
• CPC Class Codes: G06F 16/2219 Large Object storage; Management thereof
• CPC Class Codes: G06F 16/2365 Ensuring data consistency and integrity
• 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
• CPC Class Codes: G06F 16/273 Asynchronous replication or reconciliation
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9566 URL specific, e.g. using aliases, detecting broken or misspelled links
• CPC Class Codes: G06F 2201/85 Active fault masking without idle spares active fault masking without idle spare hardware where processing functionality is redundant G06F11/2035
• CPC Class Codes: G06F 3/0619 in relation to data integrity, e.g. data losses, bit errors
• CPC Class Codes: G06F 3/0626 Reducing size or complexity of storage systems
• CPC Class Codes: G06F 3/0644 Management of space entities, e.g. partitions, extents, pools
• CPC Class Codes: G06F 3/065 Replication mechanisms
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: H04L 43/106 using time related information in packets, e.g. by adding timestamps
• CPC Class Codes: H04L 65/403 Arrangements for multi-party communication, e.g. for conferences data switching systems for conference H04L12/18; arrangements for connecting several subscribers to a common circuit, i.e. affording conference facilities H04M3/56; television conferencing systems H04N7/15
• CPC Class Codes: H04L 67/06 specially adapted for file transfer, e.g. file transfer protocol [FTP]
• CPC Class Codes: H04L 67/1095 : Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes
• CPC Class Codes: H04L 67/1097 : for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• CPC Class Codes: H04L 67/5651 : Reducing the amount or size of exchanged application data
• CPC Class Codes: H04L 69/04 : Protocols for data compression, e.g. ROHC

Abstract

In one embodiment, there is a multi-cluster synchronization system between two or more clusters. The multi-cluster synchronization system uses variable compression to optimize the transfer of information between the clusters. Compression is used not only to minimize the total number of bytes sent between the two clusters, but to dynamically vary the size of the objects sent across the wire to optimize for higher throughput after considering packet loss, TCP windows, and block sizes. This includes both the packaging of multiple small files together into one larger compressed file, saving on TCP and header overhead, but also the chunking of large files into multiple smaller files that are less likely to have difficulties due to intermittent network congestion or errors. A further embodiment uses forward error correction to maximize the chances that the remote end will be able to correctly reconstitute the transmission.

Claims

1. A multi-cluster synchronization system, comprising:

an intercluster network coupling a first cluster and a remote cluster, the first cluster including a first cluster-internal network, a first structured information repository, and a first object storage, wherein the first structured information repository contains metadata corresponding to stored information objects in the first object storage, and wherein the first structured information repository and the first object storage are coupled via the cluster-internal network;

a network evaluator that determines a state of one or more networks coupled to the first cluster and the remote cluster; and

an intercluster repository synchronizer including a compression module that identifies one or more information objects to compress from the first object storage and transmit to the remote cluster in compressed form, wherein the compression module determines a target size of a single compressed information object, determines, based on the state of the one or more networks, whether to increase or decrease the target size of the single compressed information object, and updates the target size in accordance with the determining whether to increase or decrease the target size,

wherein the compression module compresses a first information object and a second information object stored in the first object storage and combines the first and second compressed information objects, wherein a size of each of the first compressed information object and second compressed information object is smaller than the updated target size, and the single information object includes the first and second compressed information objects, and

wherein the compression module transmits the single information object to the remote cluster for storage, wherein transmitting the single information object results in the duplication of the first and second information objects at the remote cluster.

2. The system of claim 1, wherein the network evaluator determines a second state of the one or more networks, and wherein the intercluster repository synchronizer dynamically determines, based on the second state of the one or more networks, a second size to compress a second information object from the first object storage, wherein the first size is different from the second size.

3. The system of claim 1, wherein the intercluster repository synchronizer manipulates the first and second information objects by compressing the first and second information objects.

4. The system of claim 3, wherein the compression module combines the first compressed information object and the second compressed information object into the single compressed information object.

5. The system of claim 3, wherein the second compressed information object is of a second size different from the first size.

6. The system of claim 2, wherein the first state of the one or more networks is different from the second state of the one or more networks.

7. The system of claim 1, wherein the state of the one or more networks is based on packet loss in the one or more networks.

8. The system of claim 1, wherein the state of the one or more networks is based on TCP windows in the one or more networks.

9. The system of claim 1, wherein the compression module determines whether the metadata in the first structured information repository corresponding to the stored information objects has changed, and in response to determining that the metadata in the first structured information repository corresponding to the stored information objects has not changed, the compression module compresses the first and second information objects into the single compressed information object.

10. The system of claim 1, wherein the compression module manipulates the information object by partitioning large data objects into a smaller size for transmission.

11. A method of synchronizing objects in an object storage system, the method comprising:

identifying a set of stored information objects stored in a first cluster to be transferred to a second cluster for storage, the first cluster and second cluster being coupled to an intercluster network;

determining a state of one or more networks coupled to the first and second clusters;

determining a target size of a single information object for transfer to the second cluster;

determining, based on the state of the one or more networks, whether to increase or decrease the target size;

updating the target size in accordance with the determining whether to increase or decrease the target size;

compressing a first information object stored in the first object storage, wherein a size of the first compressed information object is smaller than the updated target size;

compressing a second information object stored in the first object storage, wherein a size

combining the first and second compressed information objects, wherein the single information object includes the first and second compressed information objects; and

transmitting the single information object to the second cluster for storage, wherein the transmitting the single information object results in the duplication of the first and second information objects at the second cluster.

12. The method of claim 11, further including:

determining a second state of the one or more networks;

dynamically determining, based on the second state of the one or more networks, a second size to compress the second information object.

13. The method of claim 12, wherein the compressed second information object is of the second size.

14. The method of claim 13, wherein the first size is different from the second size.

15. The method of claim 12, wherein the first state of the one or more networks is different from the second state of the one or more networks.

16. The method of claim 11, wherein the state of the one or more networks is based on packet loss in the intercluster network.

17. The method of claim 11, wherein the state of the one or more networks is based on TCP windows in the intercluster network.

18. The method of claim 11, further including:

determining whether the metadata in the first structured information repository corresponding to the stored information objects has changed; and

in response to determining that the metadata in the first structured information repository corresponding to the stored information objects has not changed, compressing the first and second information objects into the single compressed information object.

19. The method of claim 11, further including:

partitioning a data object stored in the first object storage into a smaller size for transmission.

20. A non-transitory computer-readable medium comprising a plurality of machine-readable instructions that when executed by one or more processors causes the one or more processors to perform a method comprising:

identifying a set of stored information objects stored in a first cluster to be transferred to a second cluster for storage, the first cluster and second cluster being coupled to an intercluster network;

determining a state of one or more networks coupled to the first and second clusters;

determining a target size of a single information object for transfer to the second cluster;

determining, based on the state of the one or more networks, whether to increase or decrease the target size;

updating the target size in accordance with the determining whether to increase or decrease the target size;

compressing a first information object stored in the first object storage, wherein a size of the first compressed information object is smaller than the updated target size;

compressing a second information object stored in the first object storage, wherein a size of the second compressed information object is smaller than the updated target size;

combining the first and second compressed information objects, wherein the single information object includes the first and second compressed information objects; and

transmitting the single information object to the second cluster for storage, wherein the transmitting the single information object results in the duplication of the first and second information objects at the second cluster.