Integration of unstructed data into a database

• US 7,958,167 B2
• Filed: 03/05/2008
• Issued: 06/07/2011
• Est. Priority Date: 03/05/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Kamra, Vaibhav, Purtell, Michael J., Brjazovski, Arkadi, Diaconu, Cristian, Kumar, Rohan
• Primary Examiner(s): Thai; Hanh B
• Application Number:
  US12/042,571
• Publication Number:
  US 20090228429A1
• Time in Patent Office: 1,189 Days
• Field of Search: 707/104.1, 707/648, 707/791, 707/802, 707/811, 707821-826, 711/133, 711/150
• US Class Current: 707/821
• CPC Class Codes: G06F 16/1865 Transactional file systems
• CPC Class Codes: G06F 16/188 Virtual file systems

Abstract

The claimed subject matter provides a system and/or a method that extends transactional guarantees to unstructured data in a file system. A database engine, configured to retain structured data, can reserve a portion of the file system to store unstructured data to be managed by the database engine. A kernel component can associate a transaction with at least one stream that corresponds to the reserved portion of the file system. The association enables transaction based access to the reserved portion of the file system managed by the database engine.

Claims

1. In a computing environment comprising a processor operatively coupled to a memory, a computer-implemented method for facilitating transaction-based access to unstructured data stored in a file system by managing access to the unstructured data with a database engine, the method comprising:

provisioning a subset of a file system that is configured to store and manage unstructured data with one or more claim ownership blocks (COBs), the one or more COBs providing access to unstructured data stored as one or more file streams within the subset of the file system via one or more handles corresponding to the one or more file streams;

associating the one or more COBs with a database engine that is configured to store and manage structured data structured according to a data model, the one or more COBs providing the database engine access to the unstructured data;

exporting to one or more clients a set of entry points into the subset of the file system, thereby enabling local and remote access to the unstructured data corresponding to the one or more COBs via at least one share that leads into a namespace of the database engine;

the database engine creating a transaction instance that facilitates transaction-based access to the unstructured data stored within the subset of the file system;

associating at least one handle of the one or more handles and a corresponding file stream with the transaction instance; and

managing the at least one handle based at least in part on a state of the associated transaction instance to provide transactional guarantees to the corresponding file stream, the transactional guarantees including atomicity and isolation of the at least one file stream among the one or more clients.

2. The method of claim 1, further comprising flushing the at least one handle upon close of processing of the at least one handle or when the state of the transaction instance is a pre-commit state.

3. The method of claim 2, wherein flushing the at least one handle is selective in that flushing is skipped when at least one of the following conditions hold true:

file streams associated with the at least one handle have not changed;

the at least one handle is a last reference to a file stream;

or

the at least one handle references a file stream that includes a subsequent reference within a same transaction.

4. The method of claim 1, further comprising associating the at least one handle to an entry point within the set of entry points when the at least one handle originates from a remote client.

5. The method of claim 1, further comprising associating the at least one handle to a claim ownership block when the at least one handle originates from the database.

6. The method of claim 1, further comprising:

receiving a request for the file stream;

in response to the request, obtaining the at least one handle, the at least one handle corresponding to a new or existing file stream within the subset of the file system; and

determining that the request is an in-band request from the database engine, and in response, associating the at least one handle with a COB that provides access to the new or existing file stream.

7. The method of claim 6, further comprising:

modifying the new or existing file stream;

closing the new or existing file stream;

flushing the file stream; and

determining that the flush was successful, and in response, committing the transaction instance.

8. The method of claim 6, further comprising:

modifying the new or existing file stream;

closing the new or existing file stream;

flushing the file stream; and

determining that the flush was unsuccessful, and in response, rolling back the transaction instance to restore the new or existing file stream to an unmodified state.

9. The method of claim 1, further comprising:

receiving a request for the file stream;

in response to the request, obtaining the at least one handle, the at least one handle corresponding to a new or existing file stream within the subset of the file system; and

determining that the request is an out-of band request from one of the one or more clients, and in response, associating the at least one handle with a share corresponding to a COB that provides access to the new or existing file stream.

10. The method of claim 1, wherein each COB is uniquely identified by a {Volume, Path} tuple.

11. A computer-implemented system comprising a processor operatively coupled to a memory and one or more computer-readable storage media, the storage media having computer-executable instructions that, when executed by the processor, implement the method of claim 1.

12. One or more computer-readable storage media storing computer-executable instructions that, when executed by a processor, perform the method of claim 1.