Emc IP Holding Company LLC

Alexei Kabishcer

Uri Shabi

Vladimir Shveidel

Ronen Gazit

Alon Titelman

Jonathan Volij

Technology is disclosed for selectively compressing data based on similarity of pages within the data that is to be compressed. At least one corresponding hash value is generated for each one of multiple candidate pages to be compressed. In response to the hash values generated for the candidate pages, the technology selects a set of similar candidate pages from the candidate pages. The set of similar candidate pages are a subset of the candidate pages that includes less than all the candidate pages. The set of similar candidate pages are compressed as a single unit, separately from one or more other ones of the candidate pages that were not selected to be included in the set of similar candidate pages.

Selective data compression based on data similarity

Dictionary-based compression is performed to compress data units using a similar data unit as the base unit (i.e., dictionary) for each candidate data unit. Similarity may be determined between data units by applying a locality-sensitive hashing scheme to each candidate data unit to produce a hash value, and by determining whether there is a matching value in a hash index of hash values for existing data units on the system. If there is a matching hash value, the candidate data unit may be compressed using the data unit corresponding to the matching hash value as the dictionary. Only a representative portion of the data unit may be hashed to produce the hash value, the portion comprised of chunks of the data unit, where each chunk is a continuous, uninterrupted section of data. The chunks themselves may not be (in some embodiments likely are not) contiguous to one another.

Similarity deduplication

A method of operating a computing device for processing data is provided. The method includes (a) monitoring a set of performance characteristics of the processing of the data; (b) periodically calculating, using a predefined set of coefficients, a linear combination of the monitored set of performance characteristics to yield a combined metric; and (c) upon detecting that the combined metric exceeds a threshold while operating in a first processing mode, transitioning from operating in the first processing mode to operating in a second processing mode. (1) The second processing mode has a higher bandwidth than the first processing mode, and (2) processing of data in the second processing mode is less robust than processing of data in the first processing mode. An apparatus, system, and computer program product for performing a similar method are also provided.

Adaptive turbo toggling

Techniques for data processing may include: receiving a data block P having a binary representation; determining features for the data block P; determining, using at least one table of precomputed hash values, feature hashes for the features, wherein each of the feature hashes corresponds to a different feature, wherein each of the feature hashes is one of the precomputed hash values of the at least one table; and determining, in accordance with the feature hashes, a similarity hash for the data block P. Each feature may be a byte of P. The at least one table may be a single 3 dimensional or multiple 2 dimensional tables. Each row of a table of precomputed hash values may correspond to a single precomputed hash value. The row may include byte entries where each byte entry includes a single bit value of a precomputed hash.

Techniques for efficiently determining similarity hashes

A technique for managing data storage includes generating digests of chunks of data. The digests are arranged to provide similar results for similar chunks but to weight contributions from different regions of chunks differently based on their entropy. Chunks that differ in low-entropy regions tend to produce more similar digests than do chunks that differ in high-entropy regions. The technique further includes identifying similar chunks based on similarity of their digests and preforming data reduction on one chunk by reference to another, storing any differences between the chunks separately.

Facilitating data reduction using weighted similarity digest

A technique for managing data storage obtains a batch of chunks of data. The technique generates, using multiple pipelined instructions operating on the batch, a measure of data entropy for each of the chunks in the batch. The technique selectively compresses chunks in the batch based at least in part on the measures of data entropy generated for the respective chunks.

Growing community of inventors

Ramat Gan, Israel

Alexei Kabishcer