Data storage device database management architecture
A method and apparatus for a database management architecture on an SSD. A list of tables is stored in the SSD, and records of a table are stored across multiple FIMs of the SSD such that a group of records may be read in parallel by concurrently reading from multiple FIMs. The records of the table are stored on jumboblocks, organized in an unordered fashion as a linked list. New records are added to the end of the linked list. Records having gaps resulting from data modification or bad portions of an NVM die are re-organized via garbage collection when the gap memory size reaches about 20% of table memory size.
1. A data storage device, comprising:
a front-end (FE);
a front-end module (FM) comprising a plurality of flash interface modules (FIMs), a table list and a jumboblock list, the FM coupled to the FE; and
a controller comprising a processor configured to perform a method for database record management, the method comprising:
writing a first record of a table to a first location on a first FIM of the plurality of FIMs;
writing a second record of the table to a second location of a second FIM of the plurality of FIMs, wherein the second record comprises one or more gaps resulting from data modifications and wherein records of the table are stored across multiple FIMs such that a group of records can be read in parallel;
updating a first entry of the jumboblock list with the first location and a second entry of the jumboblock list with the second location; and
updating the table list with a reference to the first location of the jumboblock list.
2. The data storage device of claim 1 , wherein the first entry comprises a reference to the second entry.
3. The data storage device of claim 2 , wherein the FM further comprises a last memory buffer, the method further comprising:
receiving a third record of the table to the last memory buffer;
writing the third record to a third location on a third FIM of the plurality of FIMs;
updating a third entry of the jumboblock list with the third location; and
updating the second entry to be coupled to the third entry.
4. The data storage device of claim 3 , wherein the FM further comprises an available jumboblock list comprising at least one available location of a FIM of the plurality of FIMs, the method further comprising receiving the third location on the third FIM from the available jumboblock list.
5. The data storage device of claim 1 , wherein the first location and second location comprise sequential block addresses of the first FIM and second FIM.
6. The data storage device of claim 5 , wherein the method further comprises reading the first record and second record in parallel.
7. The data storage device of claim 6 , wherein the FM further comprises an available blocks list comprising a list of an available block of each of the plurality of FIMs.
8. The data storage device of claim 7 , wherein the method further comprises providing a third location of a third FIM for receiving a third record of the table, wherein the third location comprises a sequential block reference on the third FIM relative to the second location of the second FIM.
9. A controller for a data storage device, comprising:
an I/O to one or more memory devices; and
a processor configured to execute a method for database record management, the method comprising:
storing a first reference of a first memory location of a first flash interface module (FIM);
storing a second reference of a second memory location of a second FIM, wherein the first reference is coupled to the second reference and wherein the first and second references being coupled allow for parallel reads of memory locations of the first and second FIMs;
receiving an indication that data at one of the first memory location or second memory location has been modified, creating a gap in data; and
storing a first gap reference to the gap in data.
10. The controller of claim 9 , wherein the first reference is updated to include the first gap reference.
11. The controller of claim 10 , wherein the method further comprises receiving a request for the first and second reference.
12. The controller of claim 11 , wherein the method further comprises providing data from the first memory location, that is not referenced by the first gap reference.
13. The controller of claim 9 , further comprising a plurality of references to a plurality of memory locations of a plurality of FIMs.
14. The controller of claim 13 , wherein the first reference, second reference, and plurality of references are coupled as a linked list.
15. The controller of claim 14 , wherein at least one additional gap reference is coupled to one of the plurality of references.
16. The controller of claim 15 , wherein the method further comprises performing a garbage collection on the first gap references and the at least one additional gap reference.
17. A controller for a data storage device, comprising:
an I/O to one or more memory devices;
a processor configured to execute a method for database record management, the method comprising:
storing a first reference to a first memory location of a first flash interface module (FIM);
storing a second reference to a second memory location of a second FIM, wherein the first reference is coupled to the second reference;
receiving an indication that data at one of the first memory location and second memory location has been modified, creating a gap in data;
storing a first gap reference to the gap in data; and
performing a garbage collection on the first gap references and at least one additional gap reference, wherein the garbage collection is performed when the first gap reference and at least one additional gap reference references at least 20% of memory referenced by the first reference, second reference, and plurality of references; and
a plurality of references to a plurality of memory locations of a plurality of FIMs, wherein the first reference, second reference, and plurality of references are coupled as a linked list, and wherein the at least one additional gap reference is coupled to one of the plurality of references.
18. A system for storing data, comprising:
a front-end (FE) means;
a front-end module (FM) means comprising a table list and an available blocks list, and a data block list, the available blocks list comprising a plurality of references to a plurality of locations on one or more flash interface modules (FIMs), wherein the available blocks list allows for parallel reads of a plurality of locations on one or more FIMs; and
a controller means configured to perform a method for database record management, the method comprising:
receiving a table comprising records;
updating the table list to reference the table;
receiving a plurality of records from the table;
updating a first FIM with data from a first record of the plurality of records;
updating a second FIM with data from a second record of the plurality of records, wherein the second record contains gaps resulting from data modification; and
updating the data block list to include references to memory locations on the first and second FIMs comprising the data from the first and second records.
19. The system of claim 18 , further comprising:
updating additional FIMs with data from additional records from the plurality of records such that reading data from the first and second FIMs and additional FIMs will cause the data from the first and second records and data from additional records to be read in parallel.
20. The system of claim 18 , wherein the method further comprises providing a reference to a last buffer comprising a reference memory location in a random access memory (RAM).