CN104035822A - Low-cost efficient internal storage redundancy removing method and system - Google Patents
Low-cost efficient internal storage redundancy removing method and system Download PDFInfo
- Publication number
- CN104035822A CN104035822A CN201410231580.5A CN201410231580A CN104035822A CN 104035822 A CN104035822 A CN 104035822A CN 201410231580 A CN201410231580 A CN 201410231580A CN 104035822 A CN104035822 A CN 104035822A
- Authority
- CN
- China
- Prior art keywords
- page
- redundancy
- sampling
- memory headroom
- memory
- 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.)
- Pending
Links
Abstract
The invention discloses an internal storage space redundancy removing method and system, relates to computer internal storage managing, and belongs to the field of operation systems. The method comprises the steps that a webpage in internal storage space is sampled, wherein sampling coefficients are set, the sampling range is determined according to the sampling coefficients, and the sampled webpage is acquired according to the sampling range; the internal storage space is searched for a webpage with the webpage content same as that of the sampled webpage, the webpage serves as a redundancy webpage and is combined with the sampled webpage, and the information of the redundancy webpage is recorded; the content in the redundancy webpage is deleted, and redundancy in the internal storage space is removed. According to the internal storage space redundancy removing method and system, a scanning area can be dynamically selected, the internal storage space redundancy removing method and system are suitable for a scene with a large number of repeating webpages and also suitable for a scene with a little internal storage redundancy, and the system cost can be reduced.
Description
Technical field
The present invention relates to calculator memory management, belong to operating system field, particularly a kind of efficient internal memory de-redundancy method and system of low expense.
Background technology
Internal memory is one of of paramount importance resource of computing machine.In order to utilize efficiently limited memory source, modern operating system has been used a lot of technology to save the use of internal memory.In recent years, in (SuSE) Linux OS, there is the technology that a kind of new saving internal memory is used, i.e. internal memory de-redundancy technology, i.e. KSM (Kernel Samepage Merging, internal memory merges) technology.KSM technology is by scanning virtual memory region, find the identical physical memory page, and they are merged, internal memory during this technology makes to move and uses in the host operating system of virtual machine obviously declines, internal memory service efficiency obviously improves, thereby allows the Linux main frame of equal physical configuration can move more virtual machine after using KSM technology.
In KSM, the page is by the management of two " red-Hei " tree, and one of them " red-Hei " tree is interim, and first tree is called unstable tree, for storing also unsettled new page.In other words, as the page (content of pages does not change within a period of time) that merges candidate target, be stored in this unstable tree, the page in unstable tree is not write-protected.Second tree is called stable tree, store those and have been found that the page that is stable and merges by KSM, for determining whether a page is to stablize the page, KSM use 32 bit checks and, when a page is scanned, its verification with calculated and together with this page stores, in a follow up scan, if new verification of calculating and the verification that equals to calculate before this and, this page is to stablize the page.
While using KSM process to process a single candidate page, whether first step inspection can find the page identical with this candidate page content in stable tree, search for this stable tree, a memcmp (internal memory comparison) operation will be carried out on the page of interdependent node in this candidate page and stable tree, if memcmp returns to 0, content of pages is identical, find a matching value, otherwise, if memcmp returns to-1, represent that candidate page is less than the page of present node, if return to 1, represent that candidate page is greater than the page of present node.If the page identical with this candidate page content is arranged in stable tree, this candidate page is merged, and this candidate page is released, if do not have to find the page identical with this candidate page content in stable tree, searches for unstable tree.
When the unstable tree of search, recalculate verification in this candidate page and, if this value and original checksums and different, the subsequent searches of this scanning will be abandoned this candidate page, if verification and not change, can search for unstable tree to find the page identical with this candidate page content, if do not find the page identical with this candidate page content in unstable tree, it in unstable tree, is a new node of this candidate page interpolation, if but found the page identical with this candidate page content in unstable tree, merge this candidate page, then this node is moved in stable tree, preserve stable tree, but unstable tree is deleted and rebuild while once scanning upper.
What Recent study was more is the internal memory de-redundancy technology based on merging the same memory page, by the same memory page in seeking system, thereby these pages are merged and save Installed System Memory space, but Linux shares to improve the only system supervisor of memory efficient with the page, but its unique distinction is to implement these as operating system characteristics.VMware ESX (the fictitious host computer software of subregion and integration system under general environment) server system supervisory routine is by this characteristic called after Transparent Page Sharing (TPS, transparent page is shared), and XEN (open source code virtual machine monitor) is referred to as Memory CoW (internal memory copy-on-write, Copy on Write).No matter adopt which kind of title and implementation, this characteristic all provides better memory usage, thereby allow the excessive use internal memory of operating system (system supervisor of KVM (the system virtualization module of increasing income)), support more application program or VM (virtual machine).VMware company has early carried out the research work of virutal machine memory de-redundancy technology, the de-redundancy technology based on merging identical content memory pages that the Hypervisor (virtual machine manager) of VMware company (Vmware Inc. (US)) is used, to reduce the memory redundancy between virtual machine.This scheme is only applicable to the virtual scene of VMware, just has the effect of remarkable de-redundancy, otherwise produce little effect while there is a large amount of same memory on many virtual machines.
XEN is a popular software virtual machine, be widely used in cloud computing, VPS (Virtual Private Server, VPS) field such as, similar with KVM, XEN is also towards one of the data de-duplication technology of internal memory fabulous application scenarios, XEN Dedup (a de-redundancy system) can be good at eliminating the redundancy memory on XEN, but XEN Dedup is not based on content of pages redundancy check, the sweep limit of XEN Dedup can not be specified by user simultaneously, some enliven the page does not have the meaning merging completely, it uses scene to be just confined to the virtual scene of XEN.
Also there are the problems and shortcomings of self in KSM, lot of documents and experimental data show that it is powerful in the performance that merges redundancy memory between virtual machine, and under non-virtual scene, between general application program inside or a plurality of application programs, do not have a large amount of redundancy memories, under the considerably less scene of redundancy memory, not only to these program application KSM technology, can not reduce internal memory and use, because of the operation required internal memory of KSM itself, increase memory cost on the contrary.
This invention of patent of invention " a kind of memory redundancy removing method " relates to a kind of memory redundancy removing method, comprising: step 1, is divided into calculator memory different region units and selects; Step 2, calculates the cryptographic hash of current selection region unit; Whether step 3, search and have the cryptographic hash of other region units identical with the cryptographic hash of region unit in step 2; Step 4, if so, judges that two region units are identical as high probability, goes to step 6; Step 5, adds this region unit and cryptographic hash to search data structure, goes to step 10; Step 6, if adopt keyed Hash function in step 2, and hash function calculates is the full content of region unit, goes to step 8; Step 7, relatively whether the content of these two region units is identical; Step 8, replaces to the index of one of them region unit the index of another region unit; Step 9, discharges the region unit that is replaced index; Step 10 is selected a region unit, and is returned to step 2 from remaining area piece.But this invention is different from implementation of the present invention, wherein " a kind of memory redundancy removing method " determines the redundancy page by the mode of Hash values match, the page identical for content only retains a copy, the region page is all selected in this invention surface sweeping, and do not consider the stability of the page, as long as content of pages is identical, carry out redundancy elimination, for the page frequently writing, easily cause and need the deficiency recovered at once after elimination, and the present invention's " a kind of efficient internal memory de-redundancy method and system of low expense " designs new page scan algorithm, the all or part of selected memory page of Multiple-Scan is to find the identical redundancy page of content of pages, according to the verification of each scanning and, utilize sampled scan to realize the merging of the redundancy page to finding, after merging with two RBTree management, there is the page of different degree of stability, thereby when the elimination redundancy memory page reaches saving internal memory, improve the target of internal memory service efficiency.
Patent of invention " a kind of virtual machine (vm) migration devices and methods therefor of memory redundancy oriented ", the virtual machine (vm) migration apparatus and method of a kind of memory redundancy oriented of this disclosure of the invention, comprise: utilize data center to integrate the characteristic with internal storage data redundant virtual machine, before virtual machine formally moves, scanning obtains needs migration virutal machine memory descriptor; Descriptor memory information and the target node memory page are contrasted, obtain needing the partial memory page listings of migration; According to the memory pages list obtaining, cancel the transmission of redundancy memory, utilize data compression technique to reduce the volume of transmitted data that migration is simultaneously.Thereby, can reduce the load of network in saving the virtual machine (vm) migration process that data center's memory source is object take as far as possible, improve the resource utilization of data center's integral body, but this invention is different from implementation of the present invention, wherein " a kind of virtual machine (vm) migration devices and methods therefor of memory redundancy oriented " is by scanning, to need the judgement of migration virutal machine memory descriptor whether to have redundancy with target node memory under virtual machine (vm) migration scene, thereby in virtual machine (vm) migration process, reduce the transmission of the redundancy memory page, and the present invention is by scanning server internal memory, find out same page, for the redundancy page, only preserve a copy, thereby reducing Installed System Memory takies, can be applicable to non-virtual or virtual scene, irrelevant with virtual machine (vm) migration.
Summary of the invention
For prior art deficiency, the present invention proposes a kind of efficient internal memory de-redundancy method and system of low expense, to solve above technical matters.
The invention provides a kind of memory headroom de-redundancy method, it is characterized in that, comprising:
Step S1, samples to the page in this memory headroom, wherein sets sampling coefficient, by this sampling parameter identification sample range, and obtains the sampling page according to this sample range;
Step S2, searches the content of pages page identical with the content of this sampling page in this memory headroom, as the redundancy page, and this redundancy page and this sampling page is merged, and records the information of this redundancy page simultaneously;
Step S3, deletes the content in this redundancy page, to complete this memory headroom de-redundancy.
The efficient memory headroom de-redundancy method of described low expense, the concrete steps of step S1 comprise:
Step S11, obtains the page number m of sampling according to this sampling coefficient;
Step S12, is divided into m interval by this memory headroom;
Step S13 randomly draws a plurality of pages and forms this sampling page in each this interval.
The efficient memory headroom de-redundancy method of described low expense; this step S2 also comprises: according to the identifier of this redundancy page, search stable tree; if the identifier of this redundancy page is in this stable tree; this redundancy page and this sampling page are merged; and the identifier of this redundancy page is joined to this stable tree, this stable tree is for storing the identifier that this memory headroom has the write-protected page.
The efficient memory headroom de-redundancy method of described low expense, also comprises: this information of this redundancy page is fed back to user.
The efficient memory headroom de-redundancy method of described low expense, if this sampling coefficient reduces, this page number m reduces, otherwise increases.
The present invention also provides a kind of efficient memory headroom de-redundancy system of low expense, it is characterized in that, comprising:
Sampling module, samples for the page to this memory headroom, wherein sets sampling coefficient, by this sampling parameter identification sample range, and obtains the sampling page according to this sample range;
Merge module, for searching the page that this memory headroom content of pages is identical with the content of this sampling page, as the redundancy page, and this redundancy page and this sampling page are merged, record the information of this redundancy page simultaneously;
Removing module, for deleting the content of this redundancy page.
The efficient memory headroom de-redundancy system of described low expense, this sampling module also comprises: the page number m that obtains sampling according to this sampling coefficient; This memory headroom is divided into m interval; In each this interval, randomly draw a plurality of pages and form this sampling page.
The efficient memory headroom de-redundancy system of described low expense; also comprise and search module; for search stable tree according to the identifier of this redundancy page; if the identifier of this redundancy page is in this stable tree; this redundancy page and this sampling page are merged; and the identifier of this redundancy page is joined to this stable tree, this stable tree is for storing the identifier that this memory headroom has the write-protected page.
The efficient memory headroom de-redundancy system of described low expense, also comprises feedback module, for this information of this redundancy page is fed back to user.
The efficient memory headroom de-redundancy method of described low expense, if this sampling coefficient reduces, this page number m reduces, otherwise increases.
From above scheme, the invention has the advantages that:
The present invention can reduce the internal memory that internal memory de-redundancy algorithm itself consumes, realize a kind of new user interactions mode, can Dynamic Selection scanning area, new redundancy page quantity statistical module is proposed, thoroughly having eliminated KSM limits to originally, not only be applicable to a large amount of duplicate pages scenes, but also be applicable to a small amount of memory redundancy scene.System of the present invention is saved more internal memories than KSM, and scanning combined efficiency is higher, and system of the present invention can also be added up duplicate pages.
Accompanying drawing explanation
Fig. 1 is system module figure of the present invention;
Fig. 2 is the present invention's three DBMS structural drawing;
Fig. 3 is sampling algorithm process flow diagram of the present invention.
Wherein Reference numeral is:
The concrete steps that step 100 is sampling algorithm of the present invention, comprising:
Step 101/102/103/104/105/106/107/108/109/110/111/112/113;
Module 1 is sampling module;
Module 2 is for merging module;
Module 3 is for searching module;
Module 4 is removing module;
Module 5 is feedback module.
Embodiment
The present invention relates to internal memory de-redundancy technology in operating system, and be improvement and the innovation of on the basis of KSM, KSM page scan being carried out, overall procedure of the present invention is: the page in this memory headroom is sampled, wherein set sampling coefficient, by this sampling parameter identification sample range, and obtain the sampling page according to this sample range, wherein according to this sampling coefficient, obtain the page number m of sampling, this memory headroom is divided into m interval, in each this interval, randomly draw a plurality of pages and form this sampling page, if this sampling coefficient reduces, this page number m reduces, otherwise increase, search the content of pages page identical with the content of this sampling page in this memory headroom, as the redundancy page, and this redundancy page and this sampling page are merged, record the information of this redundancy page simultaneously, wherein according to the identifier of this redundancy page, search stable tree, if the identifier of this redundancy page is in this stable tree, this redundancy page and this sampling page are merged, and the identifier of this redundancy page is joined to this stable tree, this stable tree is for storing the identifier that this memory headroom has the write-protected page, delete the content in this redundancy page, this information of this redundancy page is fed back to user, thereby when the elimination redundancy memory page reaches saving internal memory, improved the target of internal memory service efficiency.
The scan method that the present invention proposes is that (rmap_item is a data structure according to sampling coefficient generation rmap_item, for recording the data that memory pages is relevant), when sampling coefficient is not 100, the rmap_item that the rmap_item that the scan method that the present invention proposes generates generates than KSM is few, and sampling coefficient be take certain numerical value as step-length, from 100, successively decrease, this means in most cases, scan onesize region of memory, the rmap_item generating is fewer than KSM, therefore time efficiency is better than KSM, and the impact of system performance is depended primarily on to two parameters, one is the time interval between twice sweep, another is the page quantity of each scanning, in fact these two parameters can be adjusted by sysfs file system (a kind of Virtual File System that Linux provides), the time interval between twice sweep is shorter, the page quantity of each scanning is more, larger to the performance impact of system, otherwise less.
Below in conjunction with accompanying drawing, describe the present invention, in order more clearly to describe internal memory compression method of the present invention, the explanation of given first related data structures of the present invention, the present invention adopts three DBMS structures, as shown in Figure 2, specifically comprises:
Data structure mm_slot, for recording each process virtual memory space, consists of some VMA (Virtual Memory Area), and VMA represents the virtual memory address of a certain section of centralized management of certain process; Data structure vma_node, for safeguarding the structure of VMA related data, in data structure vma_node, there is an important variable, sampling coefficient (coefficient), this coefficient is got a value n between 1 to 100, representing has the page of n% to be sampled in VMA that data structure vma_node is corresponding, if n is 100, in VMA, all pages will be sampled; For each page in the memory headroom of needs scanning generates rmap_item;
The relation of three DBMS structures is as follows: each mm_slot safeguards a RBTree, be called stable tree, store those and have been found that it is the page stable (during the double scanning page, 32 bit checks of this page and not changing, such page is called as the stable page) and that merge by KSM, in tree, linked vma_node, and vma_node has safeguarded a chained list, in chained list, linked rmap_item, i.e. mm_slot, vma_node, rmap_item tri-DBMS structures.
Be below concrete steps of the present invention, as shown in Figure 3:
Execution step 101, the rmap_item chained list gauge outfit element of current rmap_item pointed vma_node carry, execution step 102, according to sampling coefficient, by address, calculate from small to large the address of the next sampling page, concrete steps are: according to sampling coefficient calculations, go out to need the page number m of sampling, computing method are: total page number * sampling coefficient/100 of the page in memory headroom.Supposing to set initial sampling coefficient is 100, all pages of sampling in VMA, then every, reduce certain value to sampling coefficient after taking turns sampling, the page of choosing when so each sampling will be sampled than last time is few, the sampling coefficient of taking turns because former is higher, so most of duplicate pages all can be former selected in taking turns sampling, merged subsequently, formerly take turns highdensity sampling and later can enter low-density sampling, can make so again the vicissitudinous page of some contents selected union operation that carries out of having an opportunity; Virtual memory region is divided into m interval; In each interval, get at random a page.Execution step 103, whether the address that judges current rmap_item is less than the page address (address of rmap_item is less than the sampling interval inside that the page address page corresponding to explanation rmap_item sampling is positioned at this processing) sampling, if perform step 104, if the address of current rmap_item is less than the address sampling, continue judge that current rmap_item is whether in stable tree, if current rmap_item is in stable tree, perform step 106, current rmap_item points to the next page address in chained list, if current rmap_item is not in stable tree, perform step 105, from chained list, delete this rmap_item, and point to the next page address in chained list, performing step at 103 o'clock, if the address of current rmap_item is not less than the address sampling, perform step 107 and judge the address whether address of current rmap_item equals to sample, if, perform step 109, current rmap_item points to the next page in chained list, and perform step 110, whether judgement sampling finishes (whether the page number of having sampled reaches m, if reach m, illustrate that sampling finishes), if sampling does not finish, perform step 107, if sampling finishes, perform step 111, current rmap_item is carried out to the recovery logic of rmap_item to the rmap_item of chained list tail, reclaim rmap_item, the node of the rmap_item chained list that vma_node safeguards sorts from small to large according to virtual address, the operation of reclaiming rmap_item only need to travel through rmap_item chained list one time, concrete methods of realizing is: the rmap_item sampling for epicycle, the rmap_item chained list that unconditionally adds vma_node to safeguard, for on take turns the rmap_item in sampling, if found the page of the identical content that can merge, this rmap_item has been linked in stable tree, stay this rmap_item, if neither in stable tree, and by epicycle sampling, do not hit, reclaim this rmap_item.Execution step 112, sampling coefficient successively decreases, if the sampling coefficient coefficient after successively decreasing is not less than certain threshold values, returns to step 101 and carry out next round sampling, execution step 113, if sampling coefficient is less than threshold value and neither one rmap_item in stable tree, destroy vma_node, performing step at 107 o'clock, if the address of current rmap_item is not equal to the address sampling, perform step 108, create new rmap_item pointer and refer to be inserted into chained list.Above chained list is the chained list that vma_node safeguards.
The present invention has proposed a kind of efficient internal memory de-redundancy system of low expense simultaneously, comprise five modules, as shown in Figure 3: module 1 is sampling module, wherein set sampling coefficient, by this sampling parameter identification sample range, and obtain the sampling page according to this sample range; According to this sampling coefficient, obtain the page number m of sampling; This memory headroom is divided into m interval; In each this interval, randomly draw a plurality of pages and form this sampling page; According to sampling coefficient calculations sampling page address; Module 2 is for merging module, for searching the page that this memory headroom content of pages is identical with the content of this sampling page, as the redundancy page, and this redundancy page and this sampling page are merged, record the information of this redundancy page simultaneously, judge that whether the current sampling page is identical with other pages; Module 3 is for searching module, for search stable tree according to the identifier of this redundancy page, if the identifier of this redundancy page is in this stable tree, this redundancy page and this sampling page are merged, and the identifier of this redundancy page is joined to this stable tree, this stable tree is for storing the identifier that this memory headroom has the write-protected page; Redundancy page identifier is added in corresponding RBTree, and revise page pointer; Module 4 is removing module, for deleting the content of this redundancy page; Module 5 is feedback module, for this information of this redundancy page is fed back to user; Statistical redundancy page info also feeds back to user by special interface.Wherein this sampling module is set sampling coefficient, according to this sampling coefficient, obtains this sampling page, and by this sampling parameter identification sample range; If this sampling coefficient reduces, this page number m reduces, otherwise increases.
Claims (10)
1. a memory headroom de-redundancy method, is characterized in that, comprising:
Step S1, samples to the page in this memory headroom, wherein sets sampling coefficient, by this sampling parameter identification sample range, and obtains the sampling page according to this sample range;
Step S2, searches the content of pages page identical with the content of this sampling page in this memory headroom, as the redundancy page, and this redundancy page and this sampling page is merged, and records the information of this redundancy page simultaneously;
Step S3, deletes the content in this redundancy page, to complete this memory headroom de-redundancy.
2. the efficient memory headroom de-redundancy method of low expense as claimed in claim 1, is characterized in that, the concrete steps of step S1 comprise:
Step S11, obtains the page number m of sampling according to this sampling coefficient;
Step S12, is divided into m interval by this memory headroom;
Step S13 randomly draws a plurality of pages and forms this sampling page in each this interval.
3. the efficient memory headroom de-redundancy method of low expense as claimed in claim 1; it is characterized in that; this step S2 also comprises: according to the identifier of this redundancy page, search stable tree; if the identifier of this redundancy page is in this stable tree; this redundancy page and this sampling page are merged; and the identifier of this redundancy page is joined to this stable tree, this stable tree is for storing the identifier that this memory headroom has the write-protected page.
4. the efficient memory headroom de-redundancy method of low expense as claimed in claim 1, is characterized in that, also comprises: this information of this redundancy page is fed back to user.
5. the efficient memory headroom de-redundancy method of low expense as claimed in claim 1 or 2, is characterized in that, if this sampling coefficient reduces, this page number m reduces, otherwise increases.
6. an efficient memory headroom de-redundancy system for low expense, is characterized in that, comprising:
Sampling module, samples for the page to this memory headroom, wherein sets sampling coefficient, by this sampling parameter identification sample range, and obtains the sampling page according to this sample range;
Merge module, for searching the page that this memory headroom content of pages is identical with the content of this sampling page, as the redundancy page, and this redundancy page and this sampling page are merged, record the information of this redundancy page simultaneously;
Removing module, for deleting the content of this redundancy page.
7. the efficient memory headroom de-redundancy system of low expense as claimed in claim 6, is characterized in that, this sampling module also comprises: the page number m that obtains sampling according to this sampling coefficient; This memory headroom is divided into m interval; In each this interval, randomly draw a plurality of pages and form this sampling page.
8. the efficient memory headroom de-redundancy system of low expense as claimed in claim 6; it is characterized in that; also comprise and search module; for search stable tree according to the identifier of this redundancy page; if the identifier of this redundancy page is in this stable tree; this redundancy page and this sampling page are merged, and the identifier of this redundancy page is joined to this stable tree, this stable tree is for storing the identifier that this memory headroom has the write-protected page.
9. the efficient memory headroom de-redundancy system of low expense as claimed in claim 6, is characterized in that, also comprises feedback module, for this information of this redundancy page is fed back to user.
10. the efficient memory headroom de-redundancy method of the low expense as described in claim 6 or 7, is characterized in that, if this sampling coefficient reduces, this page number m reduces, otherwise increases.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410231580.5A CN104035822A (en) | 2014-05-28 | 2014-05-28 | Low-cost efficient internal storage redundancy removing method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410231580.5A CN104035822A (en) | 2014-05-28 | 2014-05-28 | Low-cost efficient internal storage redundancy removing method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104035822A true CN104035822A (en) | 2014-09-10 |
Family
ID=51466596
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410231580.5A Pending CN104035822A (en) | 2014-05-28 | 2014-05-28 | Low-cost efficient internal storage redundancy removing method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104035822A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016037344A1 (en) * | 2014-09-12 | 2016-03-17 | Intel Corporation | Memory and resource management in a virtual computing environment |
| CN107577518A (en) * | 2017-09-14 | 2018-01-12 | 郑州云海信息技术有限公司 | A kind of virtual machine deployment method and device |
| CN107924350A (en) * | 2015-08-12 | 2018-04-17 | 三星电子株式会社 | Electronic equipment and its data compression method |
| CN111352863A (en) * | 2020-03-10 | 2020-06-30 | 腾讯科技（深圳）有限公司 | Memory management method, device, equipment and storage medium |
| CN114756382A (en) * | 2022-06-14 | 2022-07-15 | 中孚安全技术有限公司 | Optimization method, system and server for memory page merging |
| WO2022247821A1 (en) * | 2021-05-26 | 2022-12-01 | 华为技术有限公司 | Memory page processing method and related device thereof |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102369520A (en) * | 2011-09-02 | 2012-03-07 | 华为技术有限公司 | Querying and traversal method and device for virtual memory area |
| CN102880555A (en) * | 2012-07-28 | 2013-01-16 | 福州大学 | Memory algorithm facing real-time system |
| CN103019884A (en) * | 2012-11-21 | 2013-04-03 | 北京航空航天大学 | Memory page de-weight method and memory page de-weight device based on virtual machine snapshot |
-
2014
- 2014-05-28 CN CN201410231580.5A patent/CN104035822A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102369520A (en) * | 2011-09-02 | 2012-03-07 | 华为技术有限公司 | Querying and traversal method and device for virtual memory area |
| CN102880555A (en) * | 2012-07-28 | 2013-01-16 | 福州大学 | Memory algorithm facing real-time system |
| CN103019884A (en) * | 2012-11-21 | 2013-04-03 | 北京航空航天大学 | Memory page de-weight method and memory page de-weight device based on virtual machine snapshot |
Non-Patent Citations (1)
| Title |
|---|
| 周霄: "Linux桌面环境下内存去重技术的研究与实现", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016037344A1 (en) * | 2014-09-12 | 2016-03-17 | Intel Corporation | Memory and resource management in a virtual computing environment |
| CN106575235A (en) * | 2014-09-12 | 2017-04-19 | 英特尔公司 | Memory and resource management in a virtual computing environment |
| US10216532B2 (en) | 2014-09-12 | 2019-02-26 | Intel Corporation | Memory and resource management in a virtual computing environment |
| CN106575235B (en) * | 2014-09-12 | 2020-10-23 | 英特尔公司 | Memory and resource management in a virtualized computing environment |
| CN107924350A (en) * | 2015-08-12 | 2018-04-17 | 三星电子株式会社 | Electronic equipment and its data compression method |
| US11226738B2 (en) | 2015-08-12 | 2022-01-18 | Samsung Electronics Co., Ltd. | Electronic device and data compression method thereof |
| CN107924350B (en) * | 2015-08-12 | 2022-01-25 | 三星电子株式会社 | Electronic equipment and data compression method thereof |
| CN107577518A (en) * | 2017-09-14 | 2018-01-12 | 郑州云海信息技术有限公司 | A kind of virtual machine deployment method and device |
| CN111352863A (en) * | 2020-03-10 | 2020-06-30 | 腾讯科技（深圳）有限公司 | Memory management method, device, equipment and storage medium |
| CN111352863B (en) * | 2020-03-10 | 2023-09-01 | 腾讯科技（深圳）有限公司 | Memory management method, device, equipment and storage medium |
| WO2022247821A1 (en) * | 2021-05-26 | 2022-12-01 | 华为技术有限公司 | Memory page processing method and related device thereof |
| CN114756382A (en) * | 2022-06-14 | 2022-07-15 | 中孚安全技术有限公司 | Optimization method, system and server for memory page merging |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11249951B2 (en) | Heuristic interface for enabling a computer device to utilize data property-based data placement inside a nonvolatile memory device | |
| CN107391653B (en) | Distributed NewSQL database system and picture data storage method | |
| US10795871B2 (en) | Key-value stores implemented using fragmented log-structured merge trees | |
| US10013317B1 (en) | Restoring a volume in a storage system | |
| US8996468B1 (en) | Block status mapping system for reducing virtual machine backup storage | |
| CN104035822A (en) | Low-cost efficient internal storage redundancy removing method and system | |
| US20130018855A1 (en) | Data deduplication | |
| JP6598996B2 (en) | Signature-based cache optimization for data preparation | |
| CN107665219B (en) | Log management method and device | |
| CN104133775A (en) | Method and apparatus for managing memory | |
| US20170364300A1 (en) | Controller, flash memory apparatus, method for identifying data block stability, and method for storing data in flash memory apparatus | |
| CN108475266B (en) | Matching fixes to remove matching documents | |
| US11100047B2 (en) | Method, device and computer program product for deleting snapshots | |
| KR102535666B1 (en) | Partitioning graph data for large scale graph processing | |
| US11262929B2 (en) | Thining databases for garbage collection | |
| CN107135662B (en) | Differential data backup method, storage system and differential data backup device | |
| CN115756312A (en) | Data access system, data access method, and storage medium | |
| CN104462388A (en) | Redundant data cleaning method based on cascade storage media | |
| JP2018530838A (en) | Cache optimization for data preparation | |
| US10416901B1 (en) | Storage element cloning in presence of data storage pre-mapper with multiple simultaneous instances of volume address using virtual copies | |
| US11347416B1 (en) | Compacting data streams in a streaming data storage platform | |
| CN105426119A (en) | Storage apparatus and data processing method | |
| EP3264254B1 (en) | System and method for a simulation of a block storage system on an object storage system | |
| US10209909B1 (en) | Storage element cloning in presence of data storage pre-mapper | |
| CN110851437A (en) | Storage method, device and equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20140910 |
|
| WD01 | Invention patent application deemed withdrawn after publication |