Method for implementing approximate data structures using operations on machine words

• US 5,519,840 A
• Filed: 01/24/1994
• Issued: 05/21/1996
• Est. Priority Date: 01/24/1994
• Status: Expired due to Fees

• Current Assignee: AT&T IPM Corp. (AT&T, Inc.)
• Original Assignee: AT&T Corporation (AT&T, Inc.)
• Inventors: Young, Neal E., Matias, Yossi, Vitter, Jeffrey S.
• Primary Examiner(s): Harrell, Robert B.
• Application Number:
  US08/187,412
• Time in Patent Office: 848 Days
• Field of Search: 364/DIG. 3 MS File, 364/DIG. 2 MS File, 395/375, 395/400, 395/425, 395/575, 395/600, 395/650, 395/800
• US Class Current: 707/791
• CPC Class Codes: G06F 9/3004 to perform operations on memory
• CPC Class Codes: Y10S 707/99942 Manipulating data structure, e.g. compression, compaction, compilation

Abstract

A method of implementing approximate data structures is disclosed. The method recognizes that an approximate data structure can be implemented using a mapping of elements in the data strucure and using an exact data strucure on the mapped elements.

Claims

1. A method of executing a first operation on a first data structure comprising a set of items to generate an output, the method comprising the steps of:

executing a second operation on machine words to generate respective outputs, the machine words representing respective second data structures comprising items, the items of each respective second data structure having a universe size no larger than the size of the machine word, each item in said set of items in said first data structure being represented by an item in at least one of said respective second data structures, and

generating said output as a function of said respective outputs.

2. The method of claim 1 wherein said second operation executed on the machine words is a successor operation.

3. The method of claim 1 wherein said second operation executed on the machine words is a predecessor operation.

4. The method of claim 1 wherein said second operation executed on the machine words is a look-up operation.

5. The method of claim 1 wherein said second operation executed on the machine words is an insertion operation.

6. The method of claim 1 wherein said second operation executed on the machine words is a deletion operation.

7. The method of claim 1 wherein a primary data structure comprises a plurality of items having an associated universe and wherein said first data structure is generated by mapping each item in said plurality of items in said primary data structure into an item in said set of items in said first data structure, said set of items in said first data structure having an associated universe smaller than the universe associated with the plurality of items in said primary data structure.

8. The method of claim 1 wherein the items in said second data structure are determined based on groups of items in said set of items in said first data structure.