Microsoft Technology Licensing, LLC

Chesong Lee

William R Tipton

Rajsekhar Das

Cornel Emilian Rat

Neal Robert Christiansen

Thomas Jay Miller

Matthew Stuart Garson

Scott Chao-Chueh Lee

Raj Das

Erik G Hortsch

Tristan Charles Griffith

Udyavara Srikanth Kamath

Erik Schmidt

Abdelkader Bahgat

Arushi Aggarwal

The use of steaming functionality on a storage device may be optimized by performing a combination of stream and non-stream writes based on a size of the data being written to a given stream. For example, a method may comprise writing data associated with a plurality of files to a first set of one or more erase blocks, determining that an amount of data associated with a given one of the plurality of files in the first set of one or more erase blocks has reached a threshold, and moving the data associated with the given file from the first set of one or more erase blocks to a stream, the stream comprising a second set of one or more erase blocks on the storage device different from the first set of one or more erase blocks.

Opportunistic use of streams for storing data on a solid state device

Embodiments described herein relate to testing the integrity of a storage system's metadata while corresponding structured data remains online. Embodiments also relate to enabling corrupt storage system metadata to be repaired while the metadata remains in use and while its structured data remains online. Corruption detection and repair is described with respect to allocation metadata and reference count metadata. The embodiments are applicable to many types of storage systems, including file systems and databases, for example.

Online repair of metadata for structured data including file systems

A file system can allocate data of a storage system into data units, wherein a set of data units can be grouped into a storage region within a volume. A process of verifying validity of data within a storage region can include obtaining a first set of error-detecting codes from a subset of the data stored in a storage system, such as from a data unit of a file system object. Each of the first set of error-detecting codes can be associated with a corresponding data unit within the storage region. A second set of error-detecting codes can be generated based at least in part on the first set of error-detecting codes, and the second set of error-detecting codes can be associated with the storage region so that the second set of error-detecting codes can be used to verify the validity of the data within the storage region.

Efficient detection of corrupt data

A file system allows a different checksum algorithm to be used for different extents of a file system object independently of other extents of the file system object. The checksum algorithm can be a selectable attribute of an extent or range of extents of a file system object, such that some extents of a file system object can use a first checksum algorithm, while other extents of the file system object can use a second checksum algorithm. An extent of the file system object also may have no associated checksum algorithm. The file system stores, for each extent of a file system object, data indicating a checksum for the extent and an indication of any checksum algorithm used for the extent.

File system with per-extent checksums

A method includes mounting a persistent volume of a data storage device of an electronic device. The persistent volume is based on a protected volume stored at the data storage device. The method also includes accessing the persistent volume to enable servicing access to the data storage device of the electronic device.

Sector-based write filtering with selective file and registry exclusions

Aspects of the subject matter described herein relate to querying and repairing data. In aspects, a component may detect that data on storage has become corrupted. In response, the component may request data from one or more redundant copies of the data and may determine which of the redundant copies, if any, are not corrupted. If a non-corrupted copy is found, the component may send a request that the corrupted data be repaired and may identify the non-corrupted copy to use to repair the corrupted data.

Growing community of inventors

Bellevue, WA, United States of America

Chesong Lee