CN100451970C - Method for realizing dynamic layout of high-performance server based on group structure - Google Patents
Method for realizing dynamic layout of high-performance server based on group structure Download PDFInfo
- Publication number
- CN100451970C CN100451970C CNB2005100448184A CN200510044818A CN100451970C CN 100451970 C CN100451970 C CN 100451970C CN B2005100448184 A CNB2005100448184 A CN B2005100448184A CN 200510044818 A CN200510044818 A CN 200510044818A CN 100451970 C CN100451970 C CN 100451970C
- Authority
- CN
- China
- Prior art keywords
- computational resource
- node
- calculating node
- storage resources
- reflection
- 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
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000005192 partition Methods 0.000 claims abstract description 5
- 230000027455 binding Effects 0.000 claims description 14
- 238000009739 binding Methods 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 14
- 230000008859 change Effects 0.000 claims description 9
- 238000000151 deposition Methods 0.000 claims description 2
- 238000010276 construction Methods 0.000 claims 1
- 230000006870 function Effects 0.000 description 40
- 238000010586 diagram Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Abstract
The present invention provides a method for realizing a dynamic layout of a high-performance server based on group structure. The present invention has the method that computing resources and storing resources of the group are divided and different identification is set. The computing resources and the storing resources provided with different identification are bound dynamically to build new computing nodes. Then the computing characteristics of the new computing nodes are changed and added to functional partitions lack of resources, and therefore, integral use ratio of the server is raised. The method for realizing the dynamic layout of the high-performance server nodes of the group structure of the present invention can ensure the computing characteristics of changing the nodes quickly and conveniently, adds idle resources of which the computing characteristic is changed dynamically to the functional partitions of which the working load is heavy, and raises the use ratio of the server.
Description
Technical field
The present invention relates to high-performance server architecture field, particularly a kind of realization of dynamic deployment method of the high-performance server based on group of planes structure.
Background technology
Along with the network technology develop rapidly, 10 Gigabit Ethernets, 10Gb infiniband network are ripe in succession to be used, and makes that the high-speed interconnect between computational resource, the storage resources becomes possibility; Along with the transfer of operating system to networked distributed system, the procotol function has become the prerequisite function of modern operating system software.In the design of parts such as computing machine network interface card, increased support to network startup, like this, in system starting process, the carrier of computing environment can be selected in certain stage, and using network storage resource to set up computing environment as carrier becomes possibility.
A group of planes is the main fluid architecture of high-performance server at present, high-performance server based on group of planes structure is by many relatively independent servers, couple together by high speed internet and to constitute, each server is called one and calculates node, each calculates on the node and all moves an independently operating system, by the cooperation and the management of software and hardware, form the high performance computer system of a single reflection.
Each calculates the hard disk that the operating system that guides on the node and move generally derives from calculating node this locality Network of Workstation.The computational resource (processor is the parts of core) and the storage resources (local hard drive is the parts of core) that calculate node are static bindings, that is to say that the estimated performance that calculates node is to be determined by operating system and related software thereof that it is stored on the local hard drive.
In the applied environment of a group of planes, a group of planes usually is deployed supports multiple application, be divided into different function divisions, at a time, the demand of using resource in the different function divisions is unbalanced, be easy to occur calculate node utilization factor height in some function division, and some subregion calculates the low phenomenon of utilization factor of node.We wish to calculate in can the dynamic adjustment function subregion quantity of node, make that calculating the nodal point number amount in the function division is complementary with the calculation task of bearing.But, because calculating the estimated performance of node is determined by the operating system and the system software of this locality, if the calculating node in the difference in functionality subregion can not change the application of being supported, even from the low function division of utilization factor, join in the high function division of utilization factor, also not necessarily can share the task in the high function division of utilization factor.
As seen, need a kind of method that can change calculating node estimated performance fast,, adjust the quantity of calculating node in the function division easily, improve the utilization factor of resource by the estimated performance of dynamic change calculating node.
Summary of the invention
The objective of the invention is to a kind of dynamic binding computational resource and storage resources, change the estimated performance that calculates node fast, improve the utilization factor of server, realize high-performance server dynamic deployment method based on group of planes structure.
Resource in the group of planes is divided into computational resource and storage resources, be respectively computational resource and different signs be set with storage resources, dynamic binding computational resource and storage resources make up new calculating node, then by changing the estimated performance that calculates node, calculating node dynamic state part after changing estimated performance is deployed in the heavy function division of operating load, to improve the overall utilization rate of server, dispositions method is divided into following steps:
A, make the operating system that in a certain specific function division, to move and the reflection of related software, and write down the application characteristic of this reflection support; Utilize the existing node (making source) in this function division to make the Image Data that need, the Image Data that copy has been made are in storage resources, record is as the base attribute of the existing node in the source of making in these Image Data, as CPU, network interface card, memory information, be used to check whether the computational resource of being bound is fit to this reflection operation.
B, search and collect initiate computational resource in the group of planes or use idle computational resource, storage resources is the storage resources on nfs server or the SAN memory device, and the binding relationship of aforementioned calculation resource and storage resources is set;
C, control computational resource start and operation from the reflection on the storage resources, discern the application program that new calculating node can be supported by the application characteristic that writes down in the reflection of depositing in the storage resources, the hardware information of automatic surveymeter operator resource, the base attribute as the existing node in the source of making of record in the hardware information that detects and the reflection is compared, determine whether computational resource meets the service condition of reflection;
The estimated performance and the resource binding of d, the new calculating node that makes up of basis concern, new node is added in the function division that needs this estimated performance goes.
Before described step b, can be to specific function zontation fixed load and time strategy, load reaches predetermined threshold values or time in this specific function subregion, automatically perform b, c, the d process is transferred to the node in the Free Partition in the heavy function division of operating load, automatically adjust the estimated performance of the calculating node in the group of planes, be fit to the applicable cases that constantly changes.
In dynamic deployment method of the present invention, computational resource is a unique identification with the MAC Address of Ethernet card of self only;
In the process of the calculating node operation SAN storage device image that has the HBA card of not supporting remote boot or HCA card, two stage bootup processs by operating system are finished, in the phase one guiding, load the driver of HBA card or HCA card to this computational resource by the Ethernet card of support remote boot in the computational resource in the new calculating node, identify the SAN memory device, in the subordinate phase guiding, the new root file system that calculates the computational resource of node is switched on the storage resources of new calculating node.
For the HBA card of not supporting remote boot or HCA card, be by the Ethernet card pilot operationp system kernel of supporting remote boot and the driver that in the computational resource of new calculating node, loads HBA card or HCA card, make it be connected SAN memory device on the network in the identification of booting operating system stage.
Description of drawings
Fig. 1 is the structural representation of computational resource and storage resources;
Fig. 2 is a group of planes structural representation;
Fig. 3 is the start-up course synoptic diagram of computational resource on NFS type stores resource;
Fig. 4 is the start-up course synoptic diagram of computational resource on SAN type stores resource.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer, below in conjunction with embodiment and accompanying drawing, the present invention is described in more detail.
The present invention has realized the dynamic combined of computational resource and storage resources in the Network of Workstation, dynamically changes the estimated performance of computational resource, and idle calculating node is added in the high function division of operating load quickly and easily, improves the utilization factor of resource.
Fig. 1 is the sign and the connection diagram of a computational resource and storage resources.Computational resource C comprises parts such as processor, internal memory, Ethernet card, HBA (HCA) card, by unique this computational resource of sign of the MAC Address of Ethernet card.
Storage resources is divided into two classes, one class is the storage resources on the nfs server, with the combination of the catalogue of nfs server IP address+NFS output as sign, note is S, another kind of resource is the resource on the SAN memory device, with SAN stored IP address+LUN number combination as sign, note is S ', Ethernet card is supported remote boot agreements such as PXE, HBA (HCA) card can not supported the remote boot agreement, the bootup process that computational resource C operates in the operating system on the SAN is after guiding successfully by Ethernet card, to switch on the SAN.
Fig. 2, be the synoptic diagram of the high-performance server of a typical group of planes structure, the computational resource C in the group of planes is because the requirement of using is dispensed in the different function divisions, be connected on the nfs server by Ethernet, be connected on the memory device by storage networking.
Usually use the pattern of a group of planes to be, deposit on nfs server or SAN and use relevant data, node is by local hard disk startup and each node of operation operating system independently.Local hard drive is to install when a group of planes is initially installed or according to needs afterwards, and the installation of operating system is a job very consuming time on the node, and the operating system after installing can not be used by other node.
Method of the present invention be by operating system installation in nfs server of sharing or SAN storage, dynamic-configuration is calculated the startup reflection of node, reaches the estimated performance of dynamic change calculating node.Calculate the reflection that node needs by quick deployment, can change the estimated performance that calculates node easily, the calculating node is joined in the function division that needs this computational resource go.
Embodiment one
As shown in Figure 2, a group of planes supports three classes to use, and the operation in function division 1,2,3 respectively of employed calculating node suppose that the operating load in the current function division 1 is very heavy, and function division 2 operating loads is very light, and dynamically the step of disposing is as follows:
1. monitoring newly adds new computational resource in the group of planes, as standby computational resource;
2. calculate the operating load of node in the monitoring group of planes in each function division, the computational resource of the lighter calculating node of operating load as standby computational resource;
3. can the computational resource of a certain calculating node in the function division 2 be extracted as C2;
4. be the reflection S1 that C2 dynamic binding function division 1 needs;
5.C2 guiding and execution reflection S1, the new calculating node C ' of structure support function subregion 1;
6. C ' is joined in the function division 1, increase the computing power of function division 1.
The prerequisite that realizes this function is, the Ethernet card network enabled of calculating node starts, and the priority that is set to network startup is higher than the startup priority of local hard drive; Application support in the function division is dynamic to be added and the deletion node.
Fig. 3 and Fig. 4 further specify and make up the concrete steps of dynamically disposing node.
Embodiment two
Shown in Figure 3 is one dynamically dispose control desk, control the process that the reflection of node from nfs server that newly joins in the group of planes starts the reciprocal process of control desk, calculating node, nfs server.As can be seen, just utilize the reflection on the nfs server to make up new calculating node, just do not need SAN equipment, the investment of SAN equipment that like this can conserve expensive if calculate node.Just nfs server can bring IO visit bottleneck as the storage of concentrating, and utilizes the method for the structure node of the memory map on the nfs server to can be regarded as a solution cheapness, that low performance requires so calculate node.
The 1-14 step is removed the 4-10 step and to set up the process based on the non-disk workstation of NFS, PXE, tftp instrument identical among Fig. 3, the purpose in 4-10 step is to check whether the computational resource and the reflection resource of binding mate, and prevents that difference because of hardware from causing the new node of structure can not normal boot and operation.
Embodiment three
As shown in Figure 4, a usefulness is dynamically disposed the reflection start-up course of node from SAN that newly joins in the group of planes of control desk control, key is two stage bootup processs of utilizing operating system itself to provide, phase one, utilize the guiding function of Ethernet card, the driver of the kernel of pilot operationp system and load store card on Ethernet (12,13 step); Subordinate phase, the storage card that does not have the network startup function is after the os starting stage finishes, root file system identifies memory device before switching, and the SAN that root file system switches to identification gone up (14 step), though this node guides on the Ethernet of low speed, but but can utilize network at a high speed in the operational process, (as infiniband, light channel network) carries out the high-speed communication and the network storage.
By the above embodiments as seen, the dynamic deployment method of the high-performance server node of this realization group of planes structure of the present invention, can guarantee fast, conveniently to change the estimated performance of node, idling-resource dynamically changed in the function division of adding to behind the estimated performance in the operating load go, improve the utilization factor of server.
Embodiment four
1) computational resource in the group of planes is separated with storage resources, the address of computational resource and associated components thereof is a unique identification with the NIC address (MAC Address) of Ethernet, and note is C; As sign, note is S with the combination of the output directory of the IP address+NFS of nfs server; As sign, note is S ' with SAN stored IP address+LUN number combination, and S and S ' are as the storage resources sign of concentrating;
2) dynamic deployment method is divided into four steps, detailed process comprises:
A, go up the operating system that making can move and the reflection of related software, and write down the application characteristic of this reflection support at S or S '
B, computational resource and storage resources binding are provided with, and the corresponding relation of C and S or S ' is set
Start c, the control reflection of C on S or S ', and with S (S ') as local storage use, constituted node in the virtual group of planes dynamically by C and S (S '), assert the application that this node can be supported by the imaged features that is connected
D, according to the characteristic of the new node that makes up this node is added in the function division of this estimated performance of needs and goes.
Wherein, the address of the link that belongs to computational resource that step 1) SAN storage switch is connected with computational resource is not as the unique identification of computational resource.
Step 2) further tells at certain hour (as the end of month of every month) according to the strategy of formulating in advance, (continue too high) under certain condition as the cpu busy percentage in certain function division, automatically perform b, c, the d process, node in the Free Partition is transferred in the heavy function division of operating load, adjusted the estimated performance of a group of planes automatically, be fit to the applicable cases that constantly changes.
Step a can further comprise the Image Data that utilize existing node (making source) to make needs; The Image Data that copy has been made; The base attribute in record making source in Image Data as CPU, network interface card, memory information, is used for checking with the computational resource binding stage whether the computational resource of being bound is fit to this reflection and moves, and sets up Image Data on the storage resources S of appointment (S ').
Step b can for: search and collect initiate computational resource in the group of planes or accept the computational resource of the not high node of utilization factor, set up the binding relationship between computational resource and the storage resources.
The startup method can further comprise the reflection of the described control computational resource of step c C on storage resources S or S ': when C starts and during operation, C directly communicates by letter with S by Ethernet card, makes up new node from S; When C starts and during operation from S ', because C can not directly communicate by letter with S ' by Ethernet card, because C is connected by HBA (Host Bus Adapter) card or HCA (HostChannel Adapter) card connection with S's ', " initialRAM disk " (or initrd) that this method adopts the Linux type operating system to provide provides two stage bootup processs, in initrd, load the driving of HBA card or HCA card, make kernel and the initrd on the NFS Server of leaving in by the guiding of MAC network interface card, can discern S ', then root file system be switched on the S '.This method also comprises the operating system (as Windows, AIX etc.) of utilizing other types, computational resource is earlier from network components guiding that can netboot, loading then can not be after network components (as the HCA card) driving of netboot, identify the computational resource S ' that directly is not connected network components that can netboot, root file system is switched to method on the S '.
Step c can for: utilize the PXE function of Ethernet card,, realize guiding and the operation of C from the S by the configuration of DHCP, tftp service; Can realize that also C goes up pilot operationp system kernel and initrd from NFS Server by configuration, utilize the driver among the initrd, discern the S ' that preliminary SAN goes up connection, the root file is switched on the S ' by DHCP, tftp service.
Steps d can be the new node that makes up, and the information that goes up reflection according to S (S ') can be known the application that new node is supported, can add new node in the function corresponding subregion to and go.
By technical scheme of the present invention as seen, the present invention is by decouples computation resource and storage resources, dynamic binding computational resource and storage resources, realized the estimated performance of node in the quick change group of planes, make the estimated performance of computational resource not be subjected to the constraint of the memory map of local hard drive, can support multiple estimated performance, according to the needs of using, reconstruct group of planes part and whole estimated performances have improved the utilization factor of computational resource and storage resources.
Claims (4)
1, a kind of dynamic deployment method of the high-performance server based on group of planes structure, it is characterized in that the resource in the group of planes is divided into computational resource and storage resources, be respectively two kinds of resources different signs is set, two kinds of resource constructions of dynamic binding become the new node that calculates, change the estimated performance that should newly calculate node then by operating system of storing on this new calculating node and software decision, and its dynamic state part is deployed in the heavy function division of operating load improves with the integral body that realizes server utilization, this dispositions method divides following steps to realize:
A, be produced on the operating system moved in a certain specific function division and the reflection of related software, and write down the application characteristic that this reflection is supported; Utilize the existing calculating node in this function division to make the Image Data that need; The Image Data that copy has been made are in storage resources; In these Image Data, also write down base attribute, be used to check whether the computational resource of being bound is fit to this reflection operation as the existing calculating node in the source of making;
B, search and collect initiate computational resource in the group of planes or use idle computational resource, storage resources is the storage resources on nfs server or the SAN memory device, the binding relationship of aforementioned calculation resource and storage resources is set, to make up new calculating node;
C, controlling described computational resource starts and operation from the reflection on the described storage resources that is held together with it, discern the application program that this new calculating node is supported by the application characteristic that writes down in the reflection of depositing in this storage resources, automatically survey the hardware information of described computational resource, the base attribute as the existing calculating node in the source of making that writes down in the hardware information that detects and the described reflection is compared, determine whether described computational resource meets the service condition of this reflection, when described computational resource meets the service condition of this reflection, continue to carry out following steps;
D, to described specific function zontation fixed load and time strategy, when load reaches predetermined threshold value in this specific function subregion or when arriving preset time, according to the estimated performance and the binding relationship of the calculating node of this new structure, this new calculating node is added in the described specific function subregion that needs its estimated performance goes; Calculating node in the Free Partition is transferred in the heavy function division of operating load, adjusted the estimated performance of the calculating node in the group of planes automatically, be fit to the applicable cases that constantly changes.
2, dynamic deployment method according to claim 1 is characterized in that, computational resource is a unique identification with the MAC Address of the Ethernet card of self.
3, dynamic deployment method according to claim 1, it is characterized in that, described calculating node is to have the HBA card of not supporting remote boot or the calculating node of HCA card, the memory device that stores map data on it is the SAN memory device, the control computational resource starts and moves from the map on the storage resources that is held together with it is to finish by two stage bootup processs of operating system, in the phase one guiding, load the driver of HBA card or HCA card to this computational resource by the Ethernet card of support remote boot in the computational resource in the new calculating node, identify the SAN memory device, in the subordinate phase guiding, the new root file system that calculates the computational resource of node is switched on the storage resources of new calculating node.
4, dynamic deployment method according to claim 3, it is characterized in that, by the Ethernet card pilot operationp system kernel of supporting remote boot and the driver that in the computational resource of new calculating node, loads HBA card or HCA card, make the HBA card of not supporting remote boot or HCA card to be connected SAN memory device on the network in the identification of booting operating system stage.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005100448184A CN100451970C (en) | 2005-09-27 | 2005-09-27 | Method for realizing dynamic layout of high-performance server based on group structure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005100448184A CN100451970C (en) | 2005-09-27 | 2005-09-27 | Method for realizing dynamic layout of high-performance server based on group structure |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1744047A CN1744047A (en) | 2006-03-08 |
| CN100451970C true CN100451970C (en) | 2009-01-14 |
Family
ID=36139434
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005100448184A Active CN100451970C (en) | 2005-09-27 | 2005-09-27 | Method for realizing dynamic layout of high-performance server based on group structure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100451970C (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9692649B2 (en) | 2014-02-26 | 2017-06-27 | International Business Machines Corporation | Role assignment for servers in a high performance computing system based on measured performance characteristics |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101820387B (en) * | 2010-02-08 | 2012-12-12 | 北京航空航天大学 | Method for rapidly deploying extensible cluster |
| CN102169448B (en) * | 2011-03-18 | 2013-10-23 | 浪潮电子信息产业股份有限公司 | Deployment method of cluster parallel computing environment |
| CN102833096B (en) * | 2012-08-06 | 2016-06-29 | 杭州华三通信技术有限公司 | The high-availability system of a kind of low cost realizes method and device |
| CN103116569A (en) * | 2012-10-31 | 2013-05-22 | 劲智数位科技股份有限公司 | Cluster type computer system with operating system environment adjustment |
| CN105703911A (en) * | 2014-11-25 | 2016-06-22 | 上海天脉聚源文化传媒有限公司 | Image processing computer and formation method |
| CN107172208B (en) * | 2017-06-30 | 2021-09-14 | 联想(北京)有限公司 | Server deployment method and system |
| CN111586144A (en) * | 2020-04-30 | 2020-08-25 | 中科院计算所西部高等技术研究院 | Computer group construction method with OODA fractal mechanism |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07250068A (en) * | 1994-03-09 | 1995-09-26 | Nippon Telegr & Teleph Corp <Ntt> | Network load smoothing method |
| CN1464405A (en) * | 2002-06-05 | 2003-12-31 | 中国科学院计算技术研究所 | A system architecture of concentration system |
| CN1512380A (en) * | 2002-12-31 | 2004-07-14 | 联想(北京)有限公司 | Load weighing method based on systematic grade diagnosis information |
| CN1547128A (en) * | 2003-12-10 | 2004-11-17 | 中国科学院计算技术研究所 | Partition lease method for cluster system resource management |
| US20040243660A1 (en) * | 2003-05-27 | 2004-12-02 | Jonathan Chew | Method and system for managing resource allocation in non-uniform resource access computer systems |
| US20050198335A1 (en) * | 2001-02-06 | 2005-09-08 | Microsoft Corporation | Distributed load balancing for single entry-point systems |
| CN1670706A (en) * | 2004-03-17 | 2005-09-21 | 联想(北京)有限公司 | A method for allocating computation nodes in cluster job management system |
-
2005
- 2005-09-27 CN CNB2005100448184A patent/CN100451970C/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07250068A (en) * | 1994-03-09 | 1995-09-26 | Nippon Telegr & Teleph Corp <Ntt> | Network load smoothing method |
| US20050198335A1 (en) * | 2001-02-06 | 2005-09-08 | Microsoft Corporation | Distributed load balancing for single entry-point systems |
| CN1464405A (en) * | 2002-06-05 | 2003-12-31 | 中国科学院计算技术研究所 | A system architecture of concentration system |
| CN1512380A (en) * | 2002-12-31 | 2004-07-14 | 联想(北京)有限公司 | Load weighing method based on systematic grade diagnosis information |
| US20040243660A1 (en) * | 2003-05-27 | 2004-12-02 | Jonathan Chew | Method and system for managing resource allocation in non-uniform resource access computer systems |
| CN1547128A (en) * | 2003-12-10 | 2004-11-17 | 中国科学院计算技术研究所 | Partition lease method for cluster system resource management |
| CN1670706A (en) * | 2004-03-17 | 2005-09-21 | 联想(北京)有限公司 | A method for allocating computation nodes in cluster job management system |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9692649B2 (en) | 2014-02-26 | 2017-06-27 | International Business Machines Corporation | Role assignment for servers in a high performance computing system based on measured performance characteristics |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1744047A (en) | 2006-03-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100451970C (en) | Method for realizing dynamic layout of high-performance server based on group structure | |
| US9712604B2 (en) | Customized configuration of cloud-based applications prior to deployment | |
| CN102521063B (en) | Shared storage method suitable for migration and fault tolerance of virtual machine | |
| US9015177B2 (en) | Dynamically splitting multi-tenant databases | |
| US20100088448A1 (en) | Virtual computing accelerator and program downloading method for server-based virtual computing | |
| US8612553B2 (en) | Method and system for dynamically purposing a computing device | |
| CN102110071B (en) | Virtual machine cluster system and implementation method thereof | |
| WO2017072933A1 (en) | Management system and management method for computer system | |
| CN104468734A (en) | Virtual cluster expanding method based on cloning | |
| US20140007092A1 (en) | Automatic transfer of workload configuration | |
| CN1848787A (en) | Automatic fast dispositioning method for aggregated server system node | |
| CN104050015A (en) | Mirror image storage and distribution system for virtual machines | |
| US11016819B2 (en) | Optimizing clustered applications in a clustered infrastructure | |
| CN111538561B (en) | OpenStack large-scale cluster deployment test method and system based on KVM virtualization technology | |
| CN103425511A (en) | System and method of installing and deploying application software in cloud computing environment | |
| CN100347672C (en) | Long-distance guide chip of transparent computing equipment based on dragon chip rack and panel construction and method thereof | |
| CN101267341A (en) | A distributed network management system, network management server and method | |
| CN109002354B (en) | OpenStack-based computing resource capacity elastic expansion method and system | |
| US20140237088A1 (en) | Managing imaging of computing devices | |
| CN113504954A (en) | Method, system and medium for calling CSI LVM plug-in, dynamic persistent volume provisioning | |
| CN102664922A (en) | High-speed network starting method based on Linux system | |
| US7668938B1 (en) | Method and system for dynamically purposing a computing device | |
| CN103577122A (en) | Method and device for achieving migration of distributed application systems between platforms | |
| CN103618634A (en) | Method for automatically finding nodes in cluster | |
| CN106911741A (en) | A kind of method and NM server for virtualizing webmaster file download load balancing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |