Data replication feedback for transport input/output

• US 9,032,032 B2
• Filed: 06/26/2008
• Issued: 05/12/2015
• Est. Priority Date: 06/26/2008
• Status: Active Grant

• Current Assignee: Zhigu Holdings Limited
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Mills, David, Luttinen, Todd, Boctor, Victor
• Primary Examiner(s): DWIVEDI, MAHESH H
• Application Number:
  US12/146,553
• Publication Number:
  US 20090327361A1
• Time in Patent Office: 2,511 Days
• Field of Search: 707/692, 709/206
• US Class Current: 709/206
• CPC Class Codes: G06F 11/1443 Transmit or communication errors
• 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 2201/835 Timestamp

Abstract

Architecture for efficiently ensuring that data is stored to the desired destination datastore such as for replication processes. A copy of data (e.g., messages) sent to a datastore for storage is stored at an alternate location until a received signal indicates that the storage and replication was successful. As soon as the feedback signal is received, the copy is removed from the alternate location, and hence, improves input/output (I/O) and storage patterns. The feedback mechanism can also be used for monitoring the status of data transport associated with log shipping, for example, and taking the appropriate actions when storage (e.g., replication) is not being performed properly.

Claims

1. A computer-implemented data management system, comprising:

a storage component to store a copy of data sent to a datastore for storage, wherein one or more parameters of the storage component is selectively configured, with the one or more parameters of the storage component to include an input/output operations per second of the storage component;

an expiration component to expire the copy at the storage component when a received signal feedback information from the datastore indicates successful storage of the data at the datastore and successful replication of the data at other datastores, the received signal to include timestamp information related to last successfully replicated data and based on a delivery time to the datastore, and the storage component to discard data fully delivered to the datastore before the timestamp to reduce input/output operations of the storage component; and

a processor to implement the storage component or the expiration component.

2. The system of claim 1, wherein the discarded data is removed from a memory cache of the storage component.

3. The system of claim 1, wherein the data is a log file of messages sent to the datastore, which is a mailbox datastore, and the copy of the log file is deleted from the storage component when the messages of the log file are successfully replicated to the mailbox datastore.

4. The system of claim 1, wherein the data is replicated to the datastore and other datastores asynchronously, and the expiration component expires the copy as soon as successful replication of the data to the datastore and the other datastores.

5. The system of claim 1, further comprising a replay component for replaying the copy in response to a failure in the storage of the data at the datastore, the expiration component expires the copy at the storage component as soon as the received signal indicates successful replay of the copy at the datastore.

6. A computer-implemented data management system, comprising:

a storage component to store a copy of messages sent to a mailbox server having a mailbox store for storage and replication, wherein one or more parameters of the storage component is selectively configured, with the one or more parameters of the storage component to include an input/output operations per second of the storage component;

a replay component to replay the copy in response to a failure in replication of the messages at the mailbox server;

an expiration component to expire the messages at the storage component when feedback information is received from the mailbox server that indicates successful replication of the messages at the mailbox server and at other mailbox stores, the feedback information to include timestamp information related to a first recipient delivery timestamp and a last recipient delivery timestamp, the storage component to discard messages that have been replicated to the mailbox server before the first recipient delivery timestamp or the last recipient delivery timestamp to reduce input/output operations of the storage component; and

a processor to implement the storage component, the replay component, or the expiration component.

7. The system of claim 6, wherein the storage component, replay component, and expiration component are part of a transport server that routes the messages to the mailbox server.

8. The system of claim 6, wherein the messages are replicated to the mailbox server and other mailbox servers asynchronously, and the expiration component expires the copy as soon as successful replication of the messages to the mailbox server and other mailbox servers occurs.

9. A computer-implemented method of managing data, comprising:

storing a copy of data, in a storage component, sent to a datastore for replication from one data store to other data stores;

configuring one or more parameters of the storage component, with the one or more parameters to include an input/output operations per second of the storage component; and

expiring the copy, based on a timestamp of a last successfully replicated file, when a received signal from the datastore indicates successful replication of the data at the datastore, by expiring messages in the copy that have timestamps before the timestamp of the last successfully replicated message to reduce input/output operations of the storage component, the timestamp based on a delivery time to the other datastores.

10. The method of claim 9, further comprising expiring the copy from a transport server cache memory.

11. The method of claim 9, further comprising communicating mailbox log replication service information as part of the received signal.

12. The method of claim 9, further comprising replaying the copy in response to a failed replication of the data at the datastore, and expiring the copy as soon as re-replication to the datastore is successful, as indicated in the received signal.

13. The method of claim 9, further comprising storing the copy in response to an asynchronous upload operation to the datastore.