Managing an analytic function to be performed on data stored in an input block

Patent metadata

Field	Value
Publication number	US-9652174-B2
Application number	US-201214398269-A
Country	US
Kind code	B2
Filing date	Jun 4, 2012
Priority date	Jun 4, 2012
Publication date	May 16, 2017
Grant date	May 16, 2017

Abstract

In an example, an analytic function to be performed on data stored in an input block is managed through an interface to a framework through which a user is to define the analytic function. The framework is to buffer batches of the data into a memory through implementation of a Reader, a Writer, a PreReader, and a PreWriter on the data stored in the input block when the user-defined analytic function is performed, and wherein the Reader, the Writer, the PreReader, and the PreWriter are individually movable with respect to each other in the input block. In addition, the user-defined analytic function is received through the interface.

First claim

What is claimed is:

1. A method implemented by a computer system comprising a physical processor that executes machine readable instructions, said method comprising:
managing, by the computer system, an analytic function to be performed on data stored in an input block by:
providing an interface to a framework through which a user is to define the analytic function to be performed on the data stored in the input block,
wherein the framework is to buffer batches of the data into a memory using at least one of a Reader, a Writer, a PreReader, and a PreWriter on the data stored in the input block responsive to the user-defined analytic function being performed, and
wherein the Reader, the Writer, the PreReader, and the PreWriter are individually movable with respect to each other in the input block; and
receiving the user-defined analytic function through the interface.

2. The method according to claim 1 , further comprising:
buffering the batches of the data into the memory in response to invoking the user-defined analytic function, without receiving instructions from the user regarding how the batches of data are to be buffered.

3. The method according to claim 2 , wherein the PreReader controls a predefined boundary of a batch of data from the Reader that is copied into the memory prior to the Reader reaching the PreReader and wherein the PreWriter controls a predefined boundary of a batch of data from the Writer that is copied into memory prior to the Writer reaching the PreWriter.

4. The method according to claim 3 , wherein the framework is to store the batch of data from the Reader in a first area of memory and store the batch of data from the Writer in a second area of memory different from the first area of memory in response to a position of the Writer being different from a position of the Reader.

5. The method according to claim 3 , further comprising:
advancing the Writer until a position of the Writer has reached a position of the PreWriter;
receiving a next call to advance the Writer;
invoking a context switch to provide control of the input block from the Writer to the PreWriter; and
moving the PreWriter to a position that defines the boundary of a next batch of data to be stored into the memory.

6. The method according to claim 5 , further comprising:
by the PreWriter,
moving the next batch of data from the input block into the memory;
copying over carry-over information to an output block;
releasing the batch of data back to an operating system; and
returning control to the Writer.

7. The method according to claim 3 , further comprising:
advancing the Reader until a position of the Reader has reached a position of the Pre Reader;
receiving a next call to advance the Reader;
invoking a context switch to provide control of the input block from the Reader to the PreReader; and
moving the PreReader to a position that defines the boundary of a next batch of data to be stored into the memory.

8. The method according to claim 7 , further comprising:
by the PreReader,
moving the next batch of data from the input block into the memory;
copying over carry-over information to an output block;
releasing the batch of data back to an operating system; and
returning control to the Reader.

9. The method according to claim 1 , wherein receiving the user-defined analytic function through the interface further comprises receiving instructions regarding when to move the Reader and the Writer during implementation of the user-defined analytic function, without receiving instructions regarding how to buffer the data.

10. An apparatus comprising:
a physical processor implementing machine readable instructions that cause the apparatus to:
manage an analytic function to be performed on data stored in an input block by:
providing an interface to a framework through which a user is to define the analytic function to be performed on the data stored in the input block,
wherein the framework is to buffer batches of the data into a memory using at least one of a Reader, a Writer, a PreReader, and a PreWriter on the data stored in the input block responsive to the user-defined analytic function being performed,
wherein the Reader, the Writer, the PreReader, and the PreWriter are individually movable with respect to each other in the input block,
wherein the PreReader controls a predefined boundary of a batch of data from the Reader that is copied into the memory prior to the Reader reaching the PreReader, and
wherein the PreWriter controls a predefined boundary of a batch of data from the Writer that is copied into memory prior to the Writer reaching the PreWriter; and

receive the user-defined analytic function through the interface.

11. The apparatus according to claim 10 , wherein the physical processor implements machine readable instructions that cause the apparatus to buffer the batch of data from the Reader and the batch of data from the Writer in response to invoking the user-defined analytic function and without receiving instructions from the user regarding how the batches of data are to be buffered.

12. The apparatus according to claim 10 , wherein the physical processor implements machine readable instructions that cause the apparatus to:
advance the Writer until a position of the Writer has reached a position of the PreWriter;
receive a next call to advance the Writer;
invoke a context switch to provide control of the input block from the Writer to the PreWriter;
move the PreWriter to a position that defines the boundary of a next batch of data to be stored into the memory;
move the next batch of data from the input block into the memory;
copy over carry-over information to an output block;
release the batch of data back to an operating system; and
return control to the Writer.

13. The apparatus according to claim 10 , wherein the physical processor implements machine readable instructions that cause the apparatus to:
advance the Reader until a position of the Reader has reached a position of the Pre Reader;
receive a next call to advance the Reader;
invoke a context switch to provide control of the input block from the Reader to the Pre Reader;
move the PreReader to a position that defines the boundary of a next batch of data to be stored into the memory;
move the next batch of data from the input block into the memory;
copy over carry-over information to an output block;
release the batch of data back to an operating system; and
return control to the Reader.

14. A non-transitory computer readable storage medium comprising machine readable instructions that when executed by a processor, cause a system to:
manage an analytic function on data stored in an input block by:
providing an interface to a framework through which a user is to define the analytic function to be performed on the data stored in the input block,
wherein the framework is to buffer batches of the data into a memory using at least one of a Reader, a Writer, a PreReader, and a PreWriter on the data stored in the input block responsive to the user-defined analytic function being performed,
wherein the Reader, the Writer, the PreReader, and the PreWriter are independently movable with respect to each other in the input block, and
wherein the PreReader controls a predefined boundary of a batch of data from the Reader that is copied into the memory prior to the Reader reaching the PreReader, and
wherein the PreWriter controls a predefined boundary of a batch of data from the Writer that is copied into memory prior to the Writer reaching the PreWriter; and

receive the user-defined analytic f

Assignees

• Chao Neil Earnest
• Fan Hongmin
• Hewlett Packard Entpr Dev Lp

Inventors

• Chao Neil Earnest
• Fan Hongmin

Classifications

• G06F12/0215

  with look ahead addressing means · CPC title

• G06F3/061 Primary

  Improving I/O performance · CPC title

• G06F12/0676

  the address being position dependent · CPC title

• G06F3/067

  Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS] · CPC title

• G06F12/0253

  Garbage collection, i.e. reclamation of unreferenced memory · CPC title