CN109714229B - Performance bottleneck positioning method of distributed storage system - Google Patents
Performance bottleneck positioning method of distributed storage system Download PDFInfo
- Publication number
- CN109714229B CN109714229B CN201811608886.2A CN201811608886A CN109714229B CN 109714229 B CN109714229 B CN 109714229B CN 201811608886 A CN201811608886 A CN 201811608886A CN 109714229 B CN109714229 B CN 109714229B
- Authority
- CN
- China
- Prior art keywords
- data
- stored
- total time
- storage system
- distributed storage
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a performance bottleneck positioning method of a distributed storage system, wherein the method comprises the following steps: s1, acquiring data to be stored; s2, storing the data to be stored to each storage node of the distributed storage system by using a distribution algorithm of the distributed storage system, and obtaining the consumed first total time; s3, performing first preset processing on each storage node, storing the data to be stored by using the processed distributed storage system, and simultaneously obtaining second total consumed time; and S4, responding to the second total time being less than the first total time, determining each storage node as a performance bottleneck, and performing performance optimization on each storage node. The method disclosed by the invention can realize the judgment of the bottleneck point of the distributed storage system.
Description
Technical Field
The invention relates to the field of performance detection, in particular to a performance bottleneck positioning method of a distributed storage system.
Background
With the continuous innovation of informatization technology and the continuous improvement of informatization level, the demands of people on the storage capacity and the computing capacity of computer equipment are explosively increased. Applications that need to store and compute TB and even PB level data do not provide fast enough computing power even with the addition of more nodes, more storage devices and processors. Therefore, efficient storage and computing power becomes a challenge that must be faced at the present time.
With the increasing importance of distributed file systems to large-scale cluster systems, governments and large computer companies have placed more emphasis on the research and investment of high-performance, scalable storage systems. Meanwhile, distributed storage systems, which are composed of inexpensive hardware and open source software, are becoming mainstream storage and computing platforms.
Generally, the structure of a distributed storage system is extremely complex, and is composed of a plurality of parts, such as a network, a back-end hard disk, a server, and the like, and any part of the whole system may become a bottleneck, so that the performance of the whole system is reduced, and the requirements of users are difficult to meet.
Disclosure of Invention
In view of the above, in order to overcome at least one aspect of the above problems, an embodiment of the present invention provides a method for locating a performance bottleneck of a distributed storage system, where the method includes:
s1, acquiring data to be stored;
s2, storing the data to be stored to each storage node of the distributed storage system by using a distribution algorithm of the distributed storage system, and obtaining the consumed first total time;
s3, performing first preset processing on each storage node, storing the storage data by using the processed distributed storage system, and obtaining second total consumed time;
and S4, responding to the second total time being less than the first total time, determining each storage node as a performance bottleneck, and performing performance optimization on each storage node.
In some embodiments, in step S3, the first preset processing is to perform virtualization processing on the storage nodes.
In some embodiments, virtualizing the respective storage nodes comprises:
constructing a virtual memory disk at each storage node;
and storing the data to be stored in the virtual memory disk.
In some embodiments, the method further comprises the steps of:
s5, responding to the second total time not less than the first total time, performing second preset processing on the distribution algorithm, storing the data to be stored by using the processed distributed storage system, and obtaining consumed third total time;
and S6, if the third total time is less than the second total time, determining that the distribution algorithm is a performance bottleneck, and performing performance optimization on the distribution algorithm.
In some embodiments, in step S5, the second preset process includes changing the distribution of the data to be stored by the distribution algorithm into the artificial distribution of the data to be stored.
In some embodiments, step S1 includes: and acquiring the data to be stored from the client by using a network card.
In some embodiments, the method further comprises the steps of:
s7, responding to the fact that the third total time is not less than the second total time, performing third preset processing on the network card, storing the data to be stored by using the processed distributed storage system, and meanwhile obtaining consumed fourth total time;
and S8, if the fourth total time is less than the third total time, determining the network card as a performance bottleneck, and performing performance optimization on the network card.
In some embodiments, in step S7, the third preset processing is to virtualize the network card.
In some embodiments, virtualizing the network card includes:
caching the data to be stored in a memory;
and storing the data to be stored cached in the memory in each storage node.
In some embodiments, virtualizing the network card further comprises:
writing a cache drive to enable the data to be stored to be cached in the memory;
writing a query driver to search the data to be stored in the memory.
The invention has the following beneficial technical effects: the performance bottleneck positioning method of the distributed storage system can realize the judgment of the bottleneck point of the distributed storage system.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other embodiments can be obtained by using the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a performance bottleneck positioning method of a distributed storage system according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a data writing flow of the Ceph distributed storage system.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the following embodiments of the present invention are described in further detail with reference to the accompanying drawings.
It should be noted that all expressions using "first" and "second" in the embodiments of the present invention are used for distinguishing two entities with the same name but different names or different parameters, and it should be noted that "first" and "second" are merely for convenience of description and should not be construed as limitations of the embodiments of the present invention, and they are not described in any more detail in the following embodiments.
According to one aspect of the invention, the embodiment of the invention provides a performance bottleneck positioning method of a distributed storage system, and the specific implementation idea is that an original distributed storage system is utilized to cache data, a total consumed time is obtained, then a key link of an IO path is resolved by analyzing and carding the data reading and writing process of the distributed storage system, a hardware device module of the key link is virtualized, the total consumed time when the virtualized distributed storage system stores data with the same size is obtained at the same time, and then the two total times are compared, so that the performance bottleneck point of the distributed storage system can be determined.
The following describes in detail a performance bottleneck positioning method of a distributed storage system according to an embodiment of the present invention with reference to fig. 1.
As shown in fig. 1, a method for locating a performance bottleneck of a distributed storage system according to an embodiment of the present invention may include:
and S1, acquiring the data to be stored.
In some embodiments, the data to be stored may be obtained directly from the local.
S2, storing the data to be stored to each storage node of the distributed storage system by using a distribution algorithm of the distributed storage system, and obtaining the consumed first total time.
In some embodiments, a first time point when the distributed storage system starts to accept the data to be stored is set, a second time point when the data to be stored is stored in each storage node is recorded, and then a first total time consumed by the distributed storage system for storing the data to be stored is obtained according to the first time point and the second time point. After the first total time is obtained, the storage speed of the distributed storage system can be obtained according to the size of the data to be stored and the first total time.
And S3, performing first preset processing on each storage node, storing the storage data by using the processed distributed storage system, and obtaining second total consumed time.
In some embodiments, after performing the first preset processing on each storage node, the data to be stored in the same size may be stored again. At this time, the time point of receiving the data to be stored is recorded again by the distributed storage system, the time point of finishing storing the data to be stored to each storage node is recorded, and then the second total time consumed by the distributed storage system for storing the data to be stored is obtained according to the two time points. After the second total time is obtained, the storage speed of the distributed storage system after the first preset processing is performed on each storage node can be obtained according to the size of the data to be stored and the second total time.
In some embodiments, the first preset processing performed on each storage node may be virtualization processing performed on each storage node, that is, virtualization of an actual hard disk device of each storage node. Specifically, a virtual memory disk may be constructed at each storage node, so that the data to be stored is stored in the virtual memory disk instead of being stored in an actual hard disk, and the influence of the read-write performance of the actual hard disk on the storage process of the distributed storage system can be shielded.
And S4, responding to the second total time being less than the first total time, determining each storage node as a performance bottleneck, and performing performance optimization on each storage node.
In some embodiments, after it is determined that the virtualization processing is performed on each storage node, the time consumed by the distributed storage system to store data of the same size is significantly reduced or the speed is significantly increased, that is, the second total time is smaller than the first total time, which indicates that the read-write speed of the actual hard disk of each storage node is a bottleneck of the system.
In some embodiments, the method for locating a performance bottleneck provided by the present invention may further include the steps of:
and S5, responding to the second total time not less than the first total time, performing second preset processing on the distribution algorithm, storing the data to be stored by using the processed distributed storage system, and obtaining consumed third total time.
In some embodiments, the second preset process may be to change the distribution of the data to be stored by a distribution algorithm into the artificial distribution of the data to be stored. The method for obtaining the third total time is the same as the method for obtaining the first total time and the second total time in the above embodiments, and details are not repeated here.
In the object storage, data is divided into a plurality of objects which are respectively stored in different storage nodes, the distribution mode is realized through a distribution algorithm in a distributed storage system, the more balanced the distribution algorithm is on the object distribution, the higher the storage efficiency of the system is, and the performance of the plurality of storage nodes can be fully utilized. The manual distribution is to manually distribute the objects to different storage nodes, so that the sufficient balanced distribution of the data can be ensured. By manually distributing the data objects, the influence of the storage system distribution algorithm on the performance due to uneven data distribution can be shielded. When the data are evenly distributed by manual means, the performance data under the optimal distribution can be obtained, the balance of the distributed storage system can be evaluated by taking the performance data as a reference, and whether the distributed algorithm is the bottleneck of the system or not is further judged.
And S6, if the third total time is less than the second total time, determining that the distribution algorithm is a performance bottleneck, and performing performance optimization on the distribution algorithm.
In some embodiments, when it is determined that the artificial distribution processing is performed, the time consumed by the distributed storage system to store data of the same size is further reduced, that is, the third total time is less than the second total time, which indicates that the distribution algorithm of the system is also a bottleneck of the system.
In some embodiments, the data to be stored may be by retrieval from a client. At this time, the system needs to receive the data to be stored from the client through the network card. The nominal rate of the network card may also be a bottleneck in the system.
If the network card is judged to be the system bottleneck, the network card can be determined to be the performance bottleneck and the performance of the network card can be optimized by performing third preset processing on the network card, storing the data to be stored by using the processed distributed storage system and obtaining the fourth total time consumed at the same time.
In some embodiments, the third preset process is to virtualize the network card. Specifically, the data to be stored is cached in a memory, and then the data to be stored cached in the memory is stored in each storage node. At this time, the data is not transmitted through the actual physical network card, but is obtained by data query cached in the memory, that is, network virtualization is realized through data retransmission, and the influence caused by network delay can be completely shielded.
In some embodiments, a time point when the distributed storage system starts to query the to-be-stored data in the memory may be set, a time point when the to-be-stored data is stored in each storage node is recorded, and then a fourth total time consumed by the distributed storage system to store the to-be-stored data may be obtained according to the two time points.
When the network card virtualization is judged to be performed, the time consumed by the distributed storage system for storing data with the same size is further reduced, namely the fourth total time is less than the third total time, which indicates that the actual physical network card of the system is also the bottleneck of the system.
In some embodiments, virtualizing the network card further comprises the steps of writing a cache driver and a query driver. The cache driver may enable the data to be stored to be cached in the memory, and the query driver may perform lookup of the data to be stored in the memory.
The following describes a performance bottleneck positioning method of a distributed storage system provided by an embodiment of the present invention by taking a Ceph distributed storage system as an example.
As shown in fig. 2, fig. 2 illustrates a write data flow of a Ceph distributed storage system.
The Ceph distributed file system mainly comprises two steps when a file writing request is carried out: firstly, a client (client) processes a file into a plurality of objects, accesses a mon node to obtain a cluster view, and sends a plurality of copies of each object to corresponding OSD (on screen display) through a cluster map and a CRUSH (hierarchical redundancy algorithm); then the OSD terminal receiving the data writing request writes the data into the actual disk. In order to ensure the safety of data, the Ceph adopts multi-copy storage, so in the data writing process, in order to ensure the consistency of multi-copy data, the Ceph firstly records related writing requests into a log and then writes into an OSD, the same data is firstly written into a main OSD and then written into a slave OSD, and the position of object storage is calculated in the same way during writing.
From the above flow analysis, it can be seen that the main links affecting the data writing performance in the distributed storage system are 4 (computing object storage location), 5 (data writing OSD), and 6(OSD copy replication). Link 4 directly affects the concurrency capability of data writing, and links 5 and 6 mainly examine the performance of the network and the hard disk, and these links are directly coupled, for example, the continuity of the calculated storage position on some OSDs directly affects the read-write response time of the hard disk, so the performance of the distributed storage system cannot be evaluated only by a single node network or the performance of the hard disk, and the actual data writing is mutually restricted in aspects.
The object storage position calculation method in the link 4 realizes the maximum equalization of data in a manual intervention mode, and can shield the influence of uneven data distribution on performance. After the influence of the shielding link 4 on the performance is shielded, the performance data obtained by the test can truly reflect the hardware configuration condition of each node in the system.
And then, the network performance is shielded through the virtualized network equipment, the virtualization of the network equipment is realized through repeated transmission of data for multiple times, a network data packet executed for the first time is firstly cached in a memory, and when the network transmission of the data packet is executed for the second time, the network data packet is directly read from the memory cache, so that a scene under the bandwidth of a virtual high-speed network is realized. After the network equipment is shielded, the performance of the storage system is not limited by the network bandwidth, and the performance result can guide a manager whether to improve the bandwidth of the network equipment.
And then shielding the rear-end hard disk by virtualizing the hard disk device, specifically, establishing an OSD service by establishing a ramdisk at each node, and eliminating the performance fault caused by the hard disk device. The test data of the rear-end hard disk can be shielded to guide whether the manager needs to speed up the rear-end hard disk.
The foregoing is an exemplary embodiment of the present disclosure, but it should be noted that various changes and modifications could be made herein without departing from the scope of the present disclosure as defined by the appended claims. The functions, steps and/or actions of the method claims in accordance with the disclosed embodiments described herein need not be performed in any particular order. Furthermore, although elements of the disclosed embodiments of the invention may be described or claimed in the singular, the plural is contemplated unless limitation to the singular is explicitly stated.
It should be understood that, as used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly supports the exception. It should also be understood that "and/or" as used herein is meant to include any and all possible combinations of one or more of the associated listed items.
The numbers of the embodiments disclosed in the embodiments of the present invention are merely for description, and do not represent the merits of the embodiments.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
Those of ordinary skill in the art will understand that: the discussion of any embodiment above is meant to be exemplary only, and is not intended to intimate that the scope of the disclosure, including the claims, of embodiments of the invention is limited to these examples; within the idea of an embodiment of the invention, also technical features in the above embodiment or in different embodiments may be combined and there are many other variations of the different aspects of an embodiment of the invention as described above, which are not provided in detail for the sake of brevity. Therefore, any omissions, modifications, substitutions, improvements, and the like that may be made without departing from the spirit and principles of the embodiments of the present invention are intended to be included within the scope of the embodiments of the present invention.
Claims (9)
1. A method for locating performance bottlenecks in a distributed storage system, wherein the method comprises the steps of:
s1, acquiring data to be stored;
s2, storing the data to be stored to each storage node of the distributed storage system by using a distribution algorithm of the distributed storage system, and obtaining the consumed first total time;
s3, performing first preset processing on each storage node, storing the storage data by using the processed distributed storage system, and obtaining second total consumed time;
s4, responding to the second total time being less than the first total time, determining each storage node as a performance bottleneck, and performing performance optimization on each storage node;
in step S3, the first preset processing is to perform virtualization processing on each storage node.
2. The method of claim 1, wherein virtualizing the storage nodes comprises:
constructing a virtual memory disk at each storage node;
and storing the data to be stored in the virtual memory disk.
3. The method of locating a performance bottleneck as recited in claim 1 wherein the method further comprises the steps of:
s5, responding to the second total time not less than the first total time, performing second preset processing on the distribution algorithm, storing the data to be stored by using the processed distributed storage system, and obtaining consumed third total time;
and S6, if the third total time is less than the second total time, determining that the distribution algorithm is a performance bottleneck, and performing performance optimization on the distribution algorithm.
4. The method according to claim 3, wherein in step S5, the second preset process includes changing the distribution of the data to be stored by a distribution algorithm into the artificial distribution of the data to be stored.
5. The method of claim 3, wherein the step S1 includes: and acquiring the data to be stored from the client by using a network card.
6. The method of locating a performance bottleneck as recited in claim 5 wherein the method further comprises the steps of:
s7, responding to the third total time not less than the second total time, performing third preset processing on the network card, storing the data to be stored by using the processed distributed storage system, and obtaining consumed fourth total time;
and S8, if the fourth total time is less than the third total time, determining the network card as a performance bottleneck, and performing performance optimization on the network card.
7. The method as claimed in claim 6, wherein in step S7, the third predetermined process is to virtualize the network card.
8. The method of claim 7, wherein virtualizing the network card comprises:
caching the data to be stored in a memory;
and storing the data to be stored cached in the memory in each storage node.
9. The method of claim 8, wherein virtualizing the network card further comprises:
writing a cache drive to enable the data to be stored to be cached in the memory;
writing a query driver to search the data to be stored in the memory.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811608886.2A CN109714229B (en) | 2018-12-27 | 2018-12-27 | Performance bottleneck positioning method of distributed storage system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811608886.2A CN109714229B (en) | 2018-12-27 | 2018-12-27 | Performance bottleneck positioning method of distributed storage system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109714229A CN109714229A (en) | 2019-05-03 |
| CN109714229B true CN109714229B (en) | 2020-09-04 |
Family
ID=66258602
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811608886.2A Active CN109714229B (en) | 2018-12-27 | 2018-12-27 | Performance bottleneck positioning method of distributed storage system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109714229B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111708488B (en) * | 2020-05-26 | 2023-01-06 | 苏州浪潮智能科技有限公司 | Distributed memory disk-based Ceph performance optimization method and device |
| CN112148219A (en) * | 2020-09-16 | 2020-12-29 | 北京优炫软件股份有限公司 | Design method and device for ceph type distributed storage cluster |
| CN112328461A (en) * | 2020-10-29 | 2021-02-05 | 无锡先进技术研究院 | Performance bottleneck prediction method, equipment and storage medium based on distributed storage |
| CN112463569B (en) * | 2020-12-11 | 2023-01-06 | 苏州浪潮智能科技有限公司 | Internal performance evaluation method and system of concurrent system |
| CN112838962B (en) * | 2020-12-31 | 2022-10-18 | ***股份有限公司 | Performance bottleneck detection method and device for big data cluster |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9106543B2 (en) * | 2013-01-22 | 2015-08-11 | International Business Machines Corporation | Signal overload management in major events and disasters |
| CN105488202A (en) * | 2015-12-09 | 2016-04-13 | 浪潮(北京)电子信息产业有限公司 | Method, apparatus and system for performance bottleneck location for distributed file system |
| CN106126407A (en) * | 2016-06-22 | 2016-11-16 | 西安交通大学 | A kind of performance monitoring Operation Optimization Systerm for distributed memory system and method |
| CN107018039A (en) * | 2016-12-16 | 2017-08-04 | 阿里巴巴集团控股有限公司 | The method and apparatus of test server clustering performance bottleneck |
| CN107145310A (en) * | 2017-05-24 | 2017-09-08 | 珠海金山网络游戏科技有限公司 | A kind of method for realizing the optimization of network storage I/O bottleneck, apparatus and system |
| CN107222331A (en) * | 2017-04-26 | 2017-09-29 | 东软集团股份有限公司 | Monitoring method, device, storage medium and the equipment of distribution application system performance |
| CN107360045A (en) * | 2017-08-31 | 2017-11-17 | 郑州云海信息技术有限公司 | The monitoring method and device of a kind of storage cluster system |
| US9836287B2 (en) * | 2013-03-21 | 2017-12-05 | Razer (Asia-Pacific) Pte. Ltd. | Storage optimization in computing devices |
| CN107911252A (en) * | 2017-12-14 | 2018-04-13 | 郑州云海信息技术有限公司 | A kind of unstructured distributed memory system method for analyzing performance, system and equipment |
| CN108959499A (en) * | 2018-06-26 | 2018-12-07 | 郑州云海信息技术有限公司 | Distributed file system performance analysis method, device, equipment and storage medium |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015195437A (en) * | 2014-03-31 | 2015-11-05 | 富士通株式会社 | Management device, information processing system, and management program |
| WO2016183544A1 (en) * | 2015-05-14 | 2016-11-17 | Walleye Software, LLC | System performance logging |
| CN106230982B (en) * | 2016-09-08 | 2019-07-16 | 哈尔滨工程大学 | A kind of dynamic self-adapting secure cloud storage method considering node reliability |
| CN107147516B (en) * | 2017-03-22 | 2020-04-28 | 华为技术有限公司 | Server, storage system and related method |
| US20180300065A1 (en) * | 2017-04-16 | 2018-10-18 | Nutanix, Inc. | Storage resource management employing end-to-end latency analytics |
-
2018
- 2018-12-27 CN CN201811608886.2A patent/CN109714229B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9106543B2 (en) * | 2013-01-22 | 2015-08-11 | International Business Machines Corporation | Signal overload management in major events and disasters |
| US9836287B2 (en) * | 2013-03-21 | 2017-12-05 | Razer (Asia-Pacific) Pte. Ltd. | Storage optimization in computing devices |
| CN105488202A (en) * | 2015-12-09 | 2016-04-13 | 浪潮(北京)电子信息产业有限公司 | Method, apparatus and system for performance bottleneck location for distributed file system |
| CN106126407A (en) * | 2016-06-22 | 2016-11-16 | 西安交通大学 | A kind of performance monitoring Operation Optimization Systerm for distributed memory system and method |
| CN107018039A (en) * | 2016-12-16 | 2017-08-04 | 阿里巴巴集团控股有限公司 | The method and apparatus of test server clustering performance bottleneck |
| CN107222331A (en) * | 2017-04-26 | 2017-09-29 | 东软集团股份有限公司 | Monitoring method, device, storage medium and the equipment of distribution application system performance |
| CN107145310A (en) * | 2017-05-24 | 2017-09-08 | 珠海金山网络游戏科技有限公司 | A kind of method for realizing the optimization of network storage I/O bottleneck, apparatus and system |
| CN107360045A (en) * | 2017-08-31 | 2017-11-17 | 郑州云海信息技术有限公司 | The monitoring method and device of a kind of storage cluster system |
| CN107911252A (en) * | 2017-12-14 | 2018-04-13 | 郑州云海信息技术有限公司 | A kind of unstructured distributed memory system method for analyzing performance, system and equipment |
| CN108959499A (en) * | 2018-06-26 | 2018-12-07 | 郑州云海信息技术有限公司 | Distributed file system performance analysis method, device, equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 分布式存储***优化技术研究;周花娜;《中国优秀硕士学位论文全文数据库》;20170315;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109714229A (en) | 2019-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109714229B (en) | Performance bottleneck positioning method of distributed storage system | |
| US10642840B1 (en) | Filtered hash table generation for performing hash joins | |
| US10545842B2 (en) | Automated local database connection affinity and failover | |
| US7680771B2 (en) | Apparatus, system, and method for database provisioning | |
| US11561930B2 (en) | Independent evictions from datastore accelerator fleet nodes | |
| US20190392047A1 (en) | Multi-table partitions in a key-value database | |
| US20150127880A1 (en) | Efficient implementations for mapreduce systems | |
| US20140115251A1 (en) | Reducing Memory Overhead of Highly Available, Distributed, In-Memory Key-Value Caches | |
| CN111708719B (en) | Computer storage acceleration method, electronic equipment and storage medium | |
| US11954118B2 (en) | Method, device and computer program product for data backup | |
| US8805797B2 (en) | Optimizing wide area network (WAN) traffic by providing home site deduplication information to a cache site | |
| CN110334145A (en) | The method and apparatus of data processing | |
| US11157456B2 (en) | Replication of data in a distributed file system using an arbiter | |
| US11288237B2 (en) | Distributed file system with thin arbiter node | |
| CN108920095B (en) | Data storage optimization method and device based on CRUSH | |
| WO2017015059A1 (en) | Efficient cache warm up based on user requests | |
| US11010091B2 (en) | Multi-tier storage | |
| US11886508B2 (en) | Adaptive tiering for database data of a replica group | |
| US20230081324A1 (en) | Shared cache for multiple index services in nonrelational databases | |
| US11860835B1 (en) | Efficient drop column requests in a non-relational data store | |
| US11126371B2 (en) | Caching file data within a clustered computing system | |
| US11609933B1 (en) | Atomic partition scheme updates to store items in partitions of a time series database | |
| US11586608B1 (en) | Handling requests to access separately stored items in a non-relational database | |
| US11765251B1 (en) | System and methods for effectively addressing fast-producer and slow-consumer problem for persistent hybrid cloud caches | |
| US11550793B1 (en) | Systems and methods for spilling data for hash joins |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |