Mechanism for data aggregation in a tracing framework

• US 8,005,794 B2
• Filed: 07/01/2008
• Issued: 08/23/2011
• Est. Priority Date: 10/31/2003
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Cantrill, Bryan M.
• Primary Examiner(s): Fleurantin; Jean B.
• Assistant Examiner(s): MYINT, DENNIS Y
• Application Number:
  US12/166,196
• Publication Number:
  US 20080263044A1
• Time in Patent Office: 1,148 Days
• Field of Search: 707/642, 707/662
• US Class Current: 707/642
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring
• CPC Class Codes: G06F 11/348 Circuit details, i.e. tracer hardware
• CPC Class Codes: G06F 11/3612 by runtime analysis performance monitoring G06F11/3466
• CPC Class Codes: G06F 11/3636 by tracing the execution of the program

Abstract

A method is disclosed for obtaining data from a kernel, including obtaining data from the kernel, and storing the data in a data set an aggregation buffer using an aggregation function. A method is disclosed for storing data in a data set, wherein the data set includes a key component, an aggregation identifier component, and a value component, including obtaining an expression, a new value, and an aggregation identifier, generating a key using the expression and the aggregation identifier; and storing the data set in a buffer, wherein storing the data set comprises storing the key in the key component, storing the aggregation identifier in the aggregation identifier component, and updating a current value in the value component using the new value and an aggregation function.

Claims

1. A method for obtaining trace data from a kernel comprising a plurality of probes, the method comprising:

obtaining first trace data from the kernel using one of the plurality of probes, wherein the first trace data comprises a first trace value and a first aggregation identifier that identifies a first aggregation function, and wherein obtaining the first trace data from the kernel comprises;

obtaining an expression, the first trace value, and the first aggregation identifier, and generating a key using the expression and the first aggregation identifier;

applying the first aggregation function to the first trace value to obtain a first aggregation result;

storing the first aggregation result in a first data set of a first aggregation buffer, wherein the first aggregation buffer is associated with a first central processing unit, wherein storing the first aggregation result in the first data set of the first aggregation buffer comprises;

storing the key in a key component of the first data set, storing the aggregation identifier in an aggregation identifier component of the first data set, and updating a value component of the first data set using the first aggregation result;

obtaining second trace data from the kernel using the one of the plurality of probes, wherein the second trace data comprises a second trace value and the first aggregation identifier;

applying the first aggregation function to the second trace value to obtain a second aggregation result;

storing the second aggregation result in a second data set of a second aggregation buffer, wherein the second aggregation buffer is associated with a second central processing unit;

applying the first aggregation function to the first aggregation result and third trace data to obtain a third aggregation result, wherein the third trace data comprises a third trace value and the first aggregation identifier;

in response to storing the third aggregation result in a user-level buffer, removing the first aggregation result from the first aggregation buffer;

applying the first aggregation function to the second aggregation result to obtain a fourth aggregation result; and

in response to storing the fourth aggregation result in the user-level buffer, removing the second aggregation result from the second aggregation buffer.

2. The method of claim 1, further comprising:

obtaining fourth trace data from the kernel using one of the plurality of probes, wherein the fourth trace data comprises a fourth trace value and a second aggregation identifier that identifies a second aggregation function;

applying the second aggregation function to the fourth trace value to obtain a fifth aggregation result;

storing the fifth aggregation result in a third data set of the first aggregation buffer;

applying the second aggregation function to the fifth aggregation result and fifth trace data to obtain a sixth aggregation result, wherein the fifth trace data comprises a fifth trace value and the second aggregation identifier; and

in response to storing the sixth aggregation result in the user-level buffer, removing the fifth aggregation result from the first aggregation buffer.

3. The method of claim 1, further comprising:

using the first aggregation identifier to obtain a user-level table key of a user-level table;

hashing the user-level table key to obtain a generated hash key;

updating a value component of a user-level table entry if a user-level table entry having a hash key matching the generated hash key is found, wherein the value component is updated using the third aggregation result;

creating a new user-level table entry if a user-level table entry having a hash key matching the generated hash key is not found, wherein creating the new user-level table entry comprises storing the generated hash value, and an initial value in a value component of the new user-level table entry in the new user-level table entry; and

updating the value component in the new user-level table entry, wherein the value component is updated using the third aggregation result.

4. The method of claim 3, wherein the user-level table is a hash table.

5. The method of claim 3, wherein the first aggregation identifier is obtained by searching at least one selected from the group consisting of a user-level dictionary and a kernel level dictionary.

6. The method of claim 1, wherein the expression comprises an n-tuple.

7. A method for obtaining trace data from a kernel comprising a plurality of probes, the method comprising:

obtaining first trace data from the kernel using one of the plurality of probes, wherein the first trace data comprises a first trace value and a first aggregation identifier that identifies a first aggregation function and wherein the first aggregation identifier is generated by applying a function to a user-defined name for an aggregation;

applying the first aggregation function to the first trace value to obtain a first aggregation result;

storing the first aggregation result in a first data set of a first aggregation buffer, wherein the first aggregation buffer is associated with a first central processing unit;

obtaining second trace data from the kernel using the one of the plurality of probes, wherein the second trace data comprises a second trace value and the first aggregation identifier;

applying the first aggregation function to the second trace value to obtain a second aggregation result;

storing the second aggregation result in a second data set of a second aggregation buffer, wherein the second aggregation buffer is associated with a second central processing unit;

applying the first aggregation function to the first aggregation result and third trace data to obtain a third aggregation result, wherein the third trace data comprises a third trace value and the first aggregation identifier;

in response to storing the third aggregation result in a user-level buffer, removing the first aggregation result from the first aggregation buffer;

applying the first aggregation function to the second aggregation result to obtain a fourth aggregation result; and

in response to storing the fourth aggregation result in the user-level buffer, removing the second aggregation result from the second aggregation buffer.

8. The method of claim 7, further comprising:

obtaining fourth trace data from the kernel using one of the plurality of probes, wherein the fourth trace data comprises a fourth trace value and a second aggregation identifier that identifies a second aggregation function;

applying the second aggregation function to the fourth trace value to obtain a fifth aggregation result;

storing the fifth aggregation result in a third data set of the first aggregation buffer;

applying the second aggregation function to the fifth aggregation result and fifth trace data to obtain a sixth aggregation result, wherein the fifth trace data comprises a fifth trace value and the second aggregation identifier; and

in response to storing the sixth aggregation result in the user-level buffer, removing the fifth aggregation result from the first aggregation buffer.

9. The method of claim 7, further comprising:

using the first aggregation identifier to obtain a user-level table key of a user-level table;

hashing the user-level table key to obtain a generated hash key;

updating a value component of a user-level table entry if a user-level table entry having a hash key matching the generated hash key is found, wherein the value component is updated using the third aggregation result;

creating a new user-level table entry if a user-level table entry having a hash key matching the generated hash key is not found, wherein creating the new user-level table entry comprises storing the generated hash value, and an initial value in a value component of the new user-level table entry in the new user-level table entry; and

updating the value component in the new user-level table entry, wherein the value component is updated using the third aggregation result.

10. The method of claim 9, wherein the user-level table is a hash table.

11. The method of claim 9, wherein the first aggregation identifier is obtained by searching at least one selected from the group consisting of a user-level dictionary and a kernel level dictionary.

12. The method of claim 7, wherein obtaining the first trace data from the kernel comprises:

obtaining an expression, the first trace value, and the first aggregation identifier, and

generating a key using the expression and the first aggregation identifier.

13. The method of claim 12, wherein storing the first aggregation result in the first data set of the first aggregation buffer comprises:

storing the key in a key component of the first data set, storing the aggregation identifier in an aggregation identifier component of the first data set, and updating a value component of the first data set using the first aggregation result.

14. The method of claim 12, wherein storing the first aggregation result in the first data set of the first aggregation buffer comprises:

generating a hash key using the key;

searching for a hash bucket corresponding to the key using the hash key;

searching for a hash chain element in the hash bucket corresponding to the key;

updating a value component of the first data set associated with the hash chain element if a hash chain element corresponding to the key is found, wherein the value component is updated using the first aggregation result;

creating a new hash chain element if the hash chain element corresponding to the key is not found, wherein creating a new hash chain element comprises associating the first data set with the new hash chain element, storing the key in a key component of the first data set, storing the aggregation identifier in an aggregation identifier component of the first data set, and storing an initial value in the value component of the first data set; and

updating the value component associated with the new hash chain element, wherein the value component is updated using the first aggregation result.

15. The method of claim 14, wherein the hash chain element is associated with the first data set using a pointer.

16. The method of claim 14, wherein the new hash chain element is associated with the first data set using a pointer.

17. The method of claim 12, wherein the expression comprises an n-tuple.