IP Library Granted Patent US 8,478,730
Granted Patent B2
US 8,478,730 · App. 13/534,361 · Granted Jul 2, 2013

Scalable deduplication system with small blocks

Inventors: Lior Aronovich (Toronto, CA); Ron Asher (Tel Aviv, IL); Michael Hirsch (Mazkeret Batya, IL); Shmuel T. Klein (Rehovot, IL); Ehud Meiri (Tel Aviv, IL); Yair Toaff (Givat Shmuel, IL)
Assignee: International Business Machines Corporation
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 8,478,730
App. No.
13/534,361
Granted
Jul 2, 2013
Kind
B2
Abstract

For scalable data deduplication working with small data chunks in a computing environment, for each of the small data chunks, a signature is generated based on a combination of a representation of characters that appear in the small data chunks with a representation of frequencies of the small data chunks. The signature is used to help in selecting the data to be deduplicated.

Claims (17)

1. A method for scalable data deduplication working with small data chunks, the method comprising:

for each of the small data chunks, generating a signature using a processing device based on a combination of a representation of characters that appear in the small data chunk and a representation of frequencies of the characters in the small data chunk, wherein the signature is used to help in selecting data to be deduplicated,

wherein the signature is generated based on a c-spectrum of the small data chunk, an f-spectrum of the small data chunk, and a p-spectrum of the small data chunk, wherein the c-spectrum of the small data chunk is a sequence of representations of different characters ordered by their frequency of occurrence in the small data chunk, the f-spectrum of the small data chunk is a corresponding sequence of frequencies of the different characters in the small data chunk, and the p-spectrum of the small data chunk is a sequence of different character pairs ordered by their frequency of occurrence in the small data chunk.

2. The method of claim 1 , wherein the signature is generated to provide for small modifications in the small data chunk to cause one of small changes or no changes in the signature.

3. The method of claim 1 , further including, in conjunction with generating the signature based on a c-spectrum, f-spectrum and p-spectrum, performing one or more of:

combining characters of the c-spectrum having similar frequencies into blocks, by weighing gaps d i =f i −f i+1 between frequencies f i of characters that are adjacent in an ordered sequence, wherein the gaps are normalized based upon weights,

selecting a subset of the gaps for defining boundaries of blocks of characters of the c-spectrum,

performing a logical XOR operation on a combination of characters in blocks of characters of the c-spectrum,

selecting a subset of bits of a subset of frequencies of the f-spectrum, and

selecting a subset of pairs of the p-spectrum and performing logical operations on pairs of representations of characters for each of the selected subset of pairs of the p-spectrum.

4. The method of claim 3 , further including performing shifting operations on the selected subset of bits and the selected subset of pairs and performing a logical XOR operation by combining the selected subsets into a bitstring representing a signature.

5. The method of claim 4 , further including performing the shifting operations according to parameter tables.

6. The method of claim 3 , wherein weights w i by which the gaps d i =f i −f i+1 are normalized are w i =1/i.

7. The method of claim 3 , wherein the logical operations performed on pairs of representations of characters include at least one of:

cyclically rotating a string of bits,

shifting the string of bits, and

performing a logical XOR operation.

Continuity (2)
Continuation 13236552 · Sep 19, 2011
Related Publication 20130073529A1 · Mar 21, 2013