DEDUPLICATION ANALYSIS
One or more aspects of the present disclosure relate to testing, analyzing, and optimizing one or more data dedup techniques implemented by a storage system. One or more workloads are generated to include zero or more deduplication (dedup) data patterns. The one or more workloads are issued to one or more storage devices. Each storage device's performance corresponding to processing the one or more workloads is analyzed.
1 . An apparatus comprising at least one processor configured to:
generate one or more workloads including zero or more deduplication (dedup) data patterns and one or more unique data patterns;
issue the one or more workloads to one or more storage devices; and
analyze each storage device's performance corresponding to processing the one or more workloads.
2 . The apparatus of claim 1 , wherein generating the one or more workloads includes generating the one or more workloads according to a target dedup ratio.
3 . The apparatus of claim 2 , wherein generating the one or more workloads includes generating zero or more duplicated patterns and one or more unique patterns, wherein each workload comprises a combination of duplicated and unique patterns.
4 . The apparatus of claim 3 , wherein the target dedup ratio corresponds to a ratio between a total number of patterns and a number of unique patterns.
5 . The apparatus of claim 3 , wherein generating either one or both of the one or more duplicated patterns and the unique patterns includes obtaining one or more of: clock timestamp, computer process identification (ID), and a shift of n-bits of a host computer number.
6 . The apparatus of claim 5 further configured to obtain the clock timestamp from an n-bit clock, wherein each clock tick of the n-bit clock corresponds to one (1) nanosecond.
7 . The apparatus of claim 3 further configured to control a granularity size of the one or more duplicated patterns.
8 . The apparatus of claim 3 further configured to generate each workload such that an order of the one or more duplicated patterns within each workload introduces a destaging delay between a first duplicated pattern and corresponding second duplicated pattern.
9 . The apparatus of claim 2 further configured to perturb the target dedup ratios of each generated workload.
10 . The apparatus of claim 1 further configured to rebalance the perturbed workloads such that each perturbed workload comprises an overall ratio of duplicated and unique patterns consistent with a target dedup ratio.
11 . A method comprising:
generating one or more workloads including zero or more deduplication (dedup) data patterns and one or more unique data patterns;
issuing the one or more workloads to one or more storage devices; and
analyzing each of the one or more storage devices performance corresponding to processing the one or more workloads.
12 . The method of claim 11 , wherein generating the one or more workloads includes generating the one or more workloads according to a target dedup ratio.
13 . The method of claim 12 , wherein generating the one or more workloads includes generating zero or more duplicated patterns and one or more unique patterns, wherein each workload comprises a combination of duplicated and unique patterns.
14 . The method of claim 13 , wherein the target dedup ratio corresponds to a ratio between a total number of patterns and a number of unique patterns.
15 . The method of claim 13 , wherein generating either one or both of the zero or more duplicated patterns and the one or more unique patterns includes obtaining one or more of: clock timestamp, computer process identification (ID), and a shift of n-bits of a host computer number.
16 . The method of claim 15 further comprising obtaining the clock timestamp from an n-bit clock, wherein each clock tick of the n-bit clock corresponds to one (1) nanosecond.
17 . The method of claim 13 further comprising controlling a granularity size of the zero or more duplicated patterns.
18 . The method of claim 13 further comprising generating each workload such that an order of the zero or more duplicated patterns within each workload introduces a destaging delay between a first duplicated pattern and corresponding second duplicated pattern.
19 . The method of claim 12 further comprising perturbing the target dedup ratios of each generated workload.
20 . The method of claim 11 further comprising rebalancing the perturbed workloads such that each perturbed workload comprises an overall ratio of duplicated and unique patterns consistent with the target dedup ratio.