REPLICATION TECHNIQUES WITH CONTENT ADDRESSABLE STORAGE

• US 20120317353A1
• Filed: 06/13/2011
• Published: 12/13/2012
• Est. Priority Date: 06/13/2011
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: XtremlO Ltd. (Dell Technologies Inc.)
• Inventors: Rokach, Ehud, WEBMAN, Erez, Frank, Shahar
• Granted Patent: US 9,383,928 B2
• Time in Patent Office: Days
• US Class Current: 711/108
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0641 De-duplication techniques
• CPC Class Codes: G06F 3/065 Replication mechanisms
• CPC Class Codes: G06F 3/0683 Plurality of storage devices

Abstract

A CAS data storage system with one or more source CAS data storage spaces and one or more destination CAS data storage spaces, and a communication line therebetween, receives input data at the source storage space for local storage and for replication to the destination CAS storage space. CAS metadata is used in the replication procedure between the two separate CAS storage spaces. Thus, data at the source storage space is used to form an active buffer for transfer to the destination storage space, the active buffer holding a hash result of the respective data item and a storage address. The system detects whenever there is more than one data item in said active buffer sharing a same storage address and upon such detection transfers a respective hash result of only the last of the data items.

Claims

1. A CAS data storage system comprising:

at least one source CAS data storage space

at least one destination CAS data storage space, and

a communication line connecting said source storage space and said destination storage space, and wherein input data for storage in said system arrives at said source storage space for storage at said source storage space and to be replicated to said destination storage space, the source storage space further comprising an active buffer of incoming data for replication to said destination storage space, said active buffer configured to hold for each of a plurality of data items of said incoming data, a hash result of the respective data item and an address, the system being configured to detect whenever there is more than one data item in said active buffer sharing a same address and upon such detection to transfer a respective hash result of only a last of said data items sharing said same address, to said destination storage space.

2. The system of claim 1, wherein said source storage space is configured to transfer said active buffer when said active buffer is complete, and said destination storage space is configured to test hash results contained in said transferred active buffer to determine whether any or all of said hash results are already present at said destination, corresponding data being requested from said source storage space only in the case of a hash result not already present at said destination storage space.

3. The system of claim 2, said destination storage space being configured to send a signal indicating whether said transfer of said corresponding data is needed, said destination storage space being further configured such that upon receipt of a hash key identical to an earlier hash key for which data transfer was required but for which said data transfer is currently incomplete, a modified acknowledge signal is issued indicating that data is awaited.

4. The system of claim 3, wherein a time-out is set following said earlier hash key, such that said modified signal indicating that said data is awaited is only sent prior to expiry of said timeout.

5. The system of claim 1, wherein said source storage space comprises a plurality of nodes each with a separate active buffer, the system being configured to complete respective active buffers simultaneously at all of said plurality of nodes and to toggle to new respective active buffers.

6. The system of claim 5, wherein said source storage space is configured to select one of said nodes as a master node to monitor said respective active buffers and to carry out said simultaneous completion and toggling.

7. The system of claim 2, wherein data items at said destination storage space are stored as data content alongside hash keys identifying said data content, said data items being de-referenced on deletion, said system being configured to reuse said de-referenced data items by comparing deleted hash keys to hash keys of incoming data items.

8. A CAS data storage system comprising at least one source CAS data storage space and a destination CAS data storage space, and a communication line connecting said source storage space and said destination storage space, and wherein input data for storage in said system arrives at said source storage space for storage in said source storage space and replication in said destination storage space, and wherein a hash key of a data item is transferred to said destination storage space, the hash key enabling the destination storage space to determine whether said data corresponding to the data item is already present at said destination storage device and does not require data transfer, said destination storage space being configured to send a signal indicating whether data transfer is required.

9. The system of claim 8, said destination storage space being further configured such that upon receipt of a hash key identical to an earlier hash key for which data transfer was required but for which said data transfer is currently incomplete a modified acknowledge signal is issued indicating that data is awaited.

10. The system of claim 8, further configured to carry out initial synchronization by sending hash keys of successively stored data items of said source storage space.

11. The system of claim 8, wherein a time-out is set following said earlier hash key, such that said modified signal indicating receipt of an identical hash key is only sent prior to expiry of said timeout.

12. A system comprising two CAS data storage spaces with synchronous replication therebetween, one of said storage spaces being a source CAS data storage space and another of said storage spaces being a destination CAS data storage space, the system further comprising a communication line connecting said source storage space and said destination storage space and having a latency, communication over said communication line comprising sending data and acknowledgements in a synchronous cycle, wherein input data for storage in said system arrives at said source storage space for storage in said source storage space and replication to said destination storage space, and wherein a hash key of a data item is transferred to said destination storage space followed within said cycle by starting to transfer corresponding data without awaiting a corresponding acknowledgement, the hash key enabling the destination storage space to determine whether said corresponding data is already present at said destination storage device and does not require data transfer, and to send an acknowledgement in said respective synchronous cycle indicating whether said data transfer is required, said acknowledgement usable at said source storage space to carry out one member of the group comprising discontinuing said started transfer of said corresponding data and acknowledging to a sending application.

13. A CAS data storage system in which data items are stored as data content alongside hash keys identifying said data content, said data items being de-referenced on deletion, said system being configured to reuse said de-referenced data items by reading said hash keys.

14. A CAS data storage method wherein input data arrives at a source CAS data storage space for storage at said source storage space and replication to a destination CAS data storage space over a communication line, the method comprising filling an active buffer, at said source storage space, with incoming data for transfer to said destination storage space, said active buffer holding for each of a plurality of data items of said incoming data, a hash result of the respective data item and a storage address, the method further comprising detecting whenever there is more than one data item in said active buffer sharing a same storage address and upon such detection transferring to said replication storage space a respective hash result of only a last of said data items sharing said same storage address.

15. The method of claim 14, further comprising:

transferring said active buffer from said source storage space to said destination storage space when said active buffer is complete,

at said destination storage space testing hash results contained in said transferred active buffer to determine whether any or all of said hash results are already present at said destination, and

requesting corresponding data from said source storage space only in the case of a hash result not already present at said destination storage space.

16. The method of claim 15, further comprising:

sending a signal from said destination storage space, indicating whether transfer of said corresponding data is required, and

upon receipt of a hash key identical to an earlier hash key for which data transfer was requested but for which said data transfer is currently incomplete a modified acknowledge signal is issued indicating that the corresponding data is awaited and should not be sent.

17. The method of claim 16, wherein a time-out is set following said earlier hash key, such that said modified signal indicating that the data is awaited is only sent prior to expiry of said timeout.

18. The method of claim 14, wherein said source storage space comprises a plurality of nodes each with a separate active buffer, the method further comprising:

completing respective active buffers simultaneously at all of said plurality of nodes; and

toggling at each of said plurality of nodes to new respective active buffers.

19. The method of claim 15, wherein data items at said destination storage space are stored as data content alongside hash keys identifying said data content, said data items being de-referenced on deletion, the method comprising reusing said de-referenced data items by comparing deleted hash keys to hash keys of incoming data items.

20. A CAS data storage method for at least one source CAS data storage space and a destination CAS data storage space, and a communication line connecting said source storage space and said destination storage space, and wherein input data for storage in said system arrives at said source storage space for storage in said destination storage space, the method comprising:

transferring a hash key of a data item to said destination storage space, the hash key enabling the destination storage space to determine whether data corresponding to the data item is already present at said destination storage device and thus does not require data transfer, and

sending a signal indicating whether said destination storage space requires data transfer.

21. The method of claim 20, further comprising,

upon receipt of a hash key identical to an earlier hash key for which data transfer was required but for which said data transfer is currently incomplete sending a modified acknowledge signal indicating that data is awaited.

22. The method of claim 21, comprising setting a time out following receipt of said earlier hash key, such that sending said modified signal indicating that data is awaited is only sent prior to expiry of said time out.

23. A synchronized CAS data storage method for at least one source CAS data storage space and a destination CAS data storage space, and a communication line connecting said source storage space and said destination storage space and having a latency, communication over said communication line comprising sending data and acknowledgements in synchronous cycles, the method comprising:

receiving input data at said source storage space for storage in said destination storage space;

transferring a hash key of a data item of said input data to said destination storage space in one of said synchronous cycles;

within said one synchronous cycle starting to transfer corresponding data without awaiting a corresponding acknowledgement, the hash key enabling the destination storage space to determine whether it already has said corresponding data and does not require data transfer;

sending an acknowledgement in said respective synchronous cycle indicating whether said data transfer is required, and

one member of the group consisting of;

discontinuing said started transfer of said corresponding data if said data transfer is not required, and sending an external acknowledgement to a source application.

24. A CAS data storage method in which data items are stored as data content alongside hash keys identifying said data content, said data items being de-referenced on deletion, the method comprising reusing said de-referenced data items by comparing hash keys corresponding to said de-referenced data to hash keys of incoming data items.

25. A first CAS storage space configured to generate and use CAS metadata for storing data items internally therein, and a second CAS storage space configured to generate and use CAS metadata for storing data items internally therein, the first CAS storage space being connected via a communication line to said second CAS storage space and configured to communicate metadata of said first CAS storage space to said second CAS storage space over said communication link in a replication procedure to replicate said first CAS storage space at said second CAS storage space, said CAS metadata comprising a hash key of a data item, the hash key corresponding to a data item to be transferred and enabling the destination storage space to determine whether data corresponding to the data item is already present at said destination storage device and thereby not requiring data transfer, said destination storage space being configured to send a signal indicating whether data transfer is required for said replication.