Write optimization extensible Hash index structure based on nonvolatile memory and insertion, refreshing and deletion methods

A non-volatile memory, hash index technology, applied in the field of write optimization and scalable hash index structure, can solve the problem of low space utilization, and achieve the goal of increasing space utilization, reducing average length, and improving data access performance. Effect

Pending Publication Date: 2021-09-03
WUHAN UNIV
View PDF4 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The present invention proposes a non-volatile memory-based write-optimized scalable hash index structure and methods for inserting, refreshing and deleting, which are used to solve or at least partially solve the technical problem of low space utilization in the methods of the prior art

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Write optimization extensible Hash index structure based on nonvolatile memory and insertion, refreshing and deletion methods
  • Write optimization extensible Hash index structure based on nonvolatile memory and insertion, refreshing and deletion methods
  • Write optimization extensible Hash index structure based on nonvolatile memory and insertion, refreshing and deletion methods

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0033] An embodiment of the present invention provides a write-optimized scalable hash index structure based on non-volatile memory, including: a pointer array, a segment, and a data bucket, wherein the pointer array includes a plurality of directory entries, and each directory entry includes a hash The hash prefix and the pointer to the corresponding segment. The hash prefix is ​​used to match the prefix of the key value calculated by the preset hash function. When the match is successful, the segment pointed to by the pointer corresponding to the hash prefix is ​​the segment to be inserted;

[0034] A segment contains multiple data buckets, and each data bucket has a serial number used to match the suffix of the hash value calculated by the preset hash function, which is the hash function of the virtual neighbor insertion algorithm;

[0035] A data bucket contains several slots, which are used to store data, and the data is key-value pairs.

[0036]Specifically, the hash ind...

Embodiment 2

[0041] Based on the same inventive concept, this application also provides a method for inserting an expandable hash index structure based on an embodiment, including:

[0042] Calculate the first hash value of the data to be inserted by using a preset hash function;

[0043] Match the prefix of the first hash value with the hash prefix stored in the directory entry in the pointer array to locate the segment to be inserted;

[0044] Determine the data bucket in the segment to be inserted according to the last byte of the first hash value, where the data bucket matching the last byte of the first hash value is used as the target data bucket, and the subsequent Adjacent data buckets are used as virtual data buckets;

[0045] Traverse the target data bucket sequentially. If there is a slot with the same key as the inserted key in the target data bucket, update the value corresponding to the current key in the slot; if there is no slot with the same key as the inserted key in the...

Embodiment 3

[0055] Based on the same inventive concept, the present application also provides a refresh method based on an extensible hash index structure in the embodiment, which adopts in-situ refresh, and during the detection process, the key value to be updated is compared with the key value stored in each slot. Key values ​​are compared, and if they are the same, the key-value pairs in the slot are updated.

[0056] In the prior art, the scalable hash structure based on NVM often uses new data to cover invalid data after the hash table is expanded, so that there will be duplicate key-value pairs, wasting space, and errors in reading data are likely to occur. The application method adopts in-situ refresh, which removes duplicate key-value pairs and improves memory utilization.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a write optimization extensible Hash index structure based on a nonvolatile memory and insertion, refreshing and deletion methods. The extensible Hash index structure comprises a pointer array, a segment, a data bucket and a slot. The insertion method comprises the steps that the Hash key finds a corresponding segment through a Virtual-Vicinity insertion algorithm by using a prefix index, then finds a data bucket of inserted data in the segment and a virtual neighborhood data bucket of the data bucket through a suffix index, and finds an idle slot to insert the data; if the insertion fails, splitting operation is performed; According to the insertion method provided by the invention, the number of insertable positions of each data can be increased, so that the space utilization rate is increased; the in-situ refreshing operation can reduce the occupation of repeated key value pairs on the memory space; in addition, the deletion operation of the HyDH can provide convenience for searching, refreshing and other operations.

Description

technical field [0001] The invention relates to the technical field of data processing, specifically designs a non-volatile storage storage structure, and in particular relates to a non-volatile memory-based write-optimized scalable hash index structure (HyDH) and methods for inserting, refreshing and deleting. Background technique [0002] Due to the advantages of high density, low power consumption and non-volatility, non-volatile memory (NVM) technology as a persistent memory is expected to be a candidate technology to supplement or replace DRAM to build future storage systems. Compared to traditional memory technologies including DRAM and SRAM. Substantial changes to low-level storage devices pose considerable challenges to the design of data structures in high-level memory and caches due to the neglect of the properties of NVM devices. A large number of existing studies have optimized tree-based index structures to efficiently fit persistent memory, such as NV-Tree, wB...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06F12/02
CPCG06F12/0238
Inventor 朱彤
Owner WUHAN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap