Patent application serial number 403891 has not been assigned to a company or institution.
The following quote was obtained by the news editors from the background information supplied by the inventors: "This invention relates to an efficient approach for utilization of checksum to support both read and write transactions. More specifically, the invention relates to a filesystem having at least two levels of data replication and utilization of the checksum technique with multiple levels of data replication.
"A checksum or hash is a fixed size datum computed from an arbitrary block of digital data for the purpose of detecting accidental errors that may have been introduced during transmission or storage. Every I/O operation on a storage device or network carries with it a small chance of introducing errors into the data that it is reading or writing. Current use of a checksum computation for a write transaction computes the checksum on a defined block size and stores the computed checksum in persistent storage and separate from the subject data. Similarly, current use of a checksum computation for a read transaction includes reading a designated block of data and computes the checksum, reading a stored checksum value, and comparing the computed checksum with the stored checksum value.
"The checksum is employed to ensure data integrity. Although there is a convenience factor associated with use of the checksum, there are performance degradation issues associated with current uses of the checksum for both read and write transactions."
In addition to the background information obtained for this patent application, VerticalNews journalists also obtained the inventors' summary information for this patent application: "This invention comprises a method, system, and article for efficient checksums for shared nothing clustered filesystems.
"In one aspect, a computer implemented method is provided for efficient checksum implementation. A write transaction is amplified to three or more levels of amplification. A first read transaction is issued to a first data block at a first replication level. The issuance of the first read transaction includes computing a first checksum on a first read block of the first read transaction. In addition, a second read checksum is issued on the first read block of the first read transaction at the second of the three or more levels of replication. Once the first and second read checksums have issued, the first and second read checksums are compared. The aspect of comparing includes serving data associated with the first transaction if the first and second read checksums match. However, if there is no match, a third read checksum is issued on the first read block of the first read transaction at a third of the three or more levels of replication.
"In another aspect, a system is provided with tools to support efficient use of checksums for shared nothing clustered filesystems. A processor is provided in communication with memory. In addition, a functional unit is provided in communication with the memory. The functional unit includes tools to support completion of a read transaction. The tools include an amplification manager, a first read manager, a second read manager, a third read manager, and a comparison manager. The amplification manager functions to amplify a write transaction. In one embodiment, the amplification manager amplifies the write transaction to three or more levels of replication. The first read manager issues a first read transaction to a first data block at a first replication level. The first read transaction includes computation of a first checksum on the first read block of the first read transaction. The second read manager computes a second read checksum on the first read block of the first read transaction at a second replication level. The comparison manager functions to compare the first and second checksums. If the first and second checksums match, the comparison manager serves data associated with the first read transaction at the first replication level. However, if the first and second checksums do not match, the third read manager computes a third read checksum on the first read block of the first read transaction at a third of the three or more levels of replication.
Most Popular Stories
- U.S. Growth Stayed Steady During Shutdown, Fed Says
- Hezbollah Chief's Assassination Claimed by Sunni Group
- Newtown Massacre Heard on 911 Recordings
- Allstate Seeks to Invest in Minority Firms
- Reid Confident Congress to Pass Immigration Bill
- Latin Music Conference Turns 25
- Guardian Pressured to Stop NSA Stories: Editor
- New Home Sales Shoot up 25 Percent in October
- Boehner Blames Obama, Senate for Congressional Inactivity
- Liberty Power Gets Minority Business Nod