Determining quality of tier assignments
AssigneeMicrosoft Technology Licensing, LLC
InventorsMikhail Bilenko, Miles Arthur Munson
Life of this patent
Estimating the term…
Abstract
Technologies pertaining to computing a tiering policy that defines how digital items are desirable stored across a plurality of different storage tiers are described herein. A data repository that comprises data that is indicative of historic user interaction with a search engine is accessed. Subsequently, a tiering policy for digital items that are retrievable by way of the search engine is computed based at least in part upon the data that is indicative of the historic user interaction with the search engine. Retrieval times for digital items in the data storage tiers differ across the data storage tiers.
Claims19
A method, comprising:
- accessing a data repository that comprises data that is indicative of historic user interaction with a search engine, the data that is indicative of the historic user interaction with the search engine comprising a plurality of queries issued by users of the search engine;
- causing a processor to compute a value that is indicative of quality of a first tiering assignment of digital items that are retrievable by way of the search engine, the first tiering assignment being an assignment of the digital items across respective data storage tiers that are accessible by the search engine, the value that is indicative of quality of the first tiering assignment computed based at least in part upon the data that is indicative of historic user interaction with the search engine, the first tiering assignment, and probabilities that an nth tier in the data storage tiers will be reached when conducting searches using the plurality of queries;
- and
- causing the processor to compute a tiering policy for the digital items based at least in part upon the value that is indicative of the quality of the first tiering assignment, the tiering policy defining how the digital items are selectively stored across the data storage tiers that are accessible to the search engine, wherein retrieval times for digital items in the data storage tiers differ across the data storage tiers.
A system, comprising:
- a processor;
- and
- memory that comprises a plurality of components that are executable by the processor, the plurality of components comprising:
- a receiver component that is configured to receive:
- user history data with respect to a search engine, the user history data comprises a plurality of queries issued by users of a search engine;
- and
- probabilities that an nth tier in a plurality of data storage tiers will be reached by the search engine when conducting searched based upon the plurality of queries;
- a quality indicator component that is configured to compute a value that is indicative of quality of a first tier assignment based upon the user history data and the probabilities, the first tier assignment being an assignment of a plurality of digital items retrievable by the search engine across the plurality of data storage tiers, wherein a retrieval time of a first storage tier in the plurality of data storage tiers is faster than a retrieval time of a second storage tier in the plurality of data storage tiers;
- and
- an update component that is configured to compute a tiering policy for the plurality of digital items based at least in part upon the value that is indicative of quality of the first tier assignment, the tiering policy defines where the plurality of digital items are to be selectively stored across the plurality of data storage tiers.
A computer-readable data storage device comprising instructions that, when executed by a processor, cause the processor to perform acts comprising:
- receiving data that is indicative of historic user interaction with a search engine, the search engine comprising a search engine index that is spread across a plurality of data storage tiers in accordance with a first tier assignment, the first tier assignment being an assignment of a plurality of digital items in the search engine index across the plurality of data storage tiers such that each data storage tier includes respective digital items, the first tier assignment is conducted in accordance with a first tiering policy that defines where, across the plurality of data storage tiers, the plurality of digital items are to be stored, data retrieval speeds being different for each storage tier in the plurality of data storage tiers;
- computing a value that is indicative of quality of the first tier assignment based at least in part upon:
- the data that is indicative of historic user interaction with the search engine;
- and
- probabilities that an nth storage tier in the plurality of data storage tiers will be reached by the search engine when conducting respective searches using the queries;
- and
- computing a second tiering policy that defines where, across the plurality of data storage tiers, the plurality of digital items in the search engine index are to be stored, the second tiering policy computed based upon the value that is indicative of quality of the first tier assignment, wherein the second tiering policy indicates that a digital item in the plurality of digital items is to be reassigned from a first data storage tier to a second data storage tier, the digital item stored in the first tier in the first tier assignment.
Dependencies are inferred from the claim text.