CN100422937C

CN100422937C - Grid organization

Info

Publication number: CN100422937C
Application number: CNB2004800216405A
Authority: CN
Inventors: 埃罗尔·博扎克; 亚历山大·格布哈特
Original assignee: SAP SE
Current assignee: SAP SE
Priority date: 2003-07-28
Filing date: 2004-07-28
Publication date: 2008-10-01
Anticipated expiration: 2024-07-28
Also published as: CN1829968A; CN1829970B; CN1829963A; CN1829966A; CN100378669C; CN100440152C; CN100472455C; CN1829969A; CN1829970A; CN1829965A; CN100422938C; CN1829967A; CN1829963B

Abstract

A method includes, in a grid computing environment, maintaining systems having grid managers having hierarchical relations, the relations of each grid manager stored in each of the systems, Each of these hierarchical relations are classified as superior or inferior.

Description

System and maintaining method, the network system of grid managers arranged and handle service method

The cross reference of related application

It is EV331001684US, title for " GRID COMPUTING MANAGEMENT " that the application's mode by reference combines piece number that submitted on July 28th, 2003, fast, give the content of people's such as ErolBozak No. the 60/490th, 818, U.S. Provisional Application.

Technical field

The present invention relates to data processing, and more specifically, relate to the Dynamic Tree Structure that is used for grid computing by the digital machine execution.

Background technology

In the data center of today, the cluster of server in the commercial client-server network of using of operation is often being performed poor aspect the unpredictable working load of management.A server may be in idle condition, and another is compeled for work. this has caused following awkward situation: need avoid the company of network bottleneck and assurance and client, business parnter and employee's connectedness to lay a course at the peak-peak in the workload demands usually, notice then: those superfluous server mosts of the time underload ground operational excellence.

In grid computing, all heterogeneous computing machines in a mechanism or between a plurality of mechanisms and system have become a large-scale integrated computing system. and then, this single system ensemble can be handled all too big and too intensive and problem and processing that can not handle easily for any single computing machine with effective and efficient manner.

More particularly, grid computing is a kind of distributed system of form, and wherein spanning network is shared computational resource.Grid computing makes it possible to select, concentrates and share and resides in a plurality of management domains and cross-regional information resources.For example availability, ability and cost and the quality of services for users (QoS) based on these information resources requires to share these information resources.Grid computing may mean: reduce the entitlement cost, and centralized calculation, data and storage resources and the efficient that improves calculating, data and storage resources, and make it possible to Virtual Organization's application and data sharing.

Summary of the invention

In one aspect, the present invention has the system of grid managers with a kind of maintenance method is a feature, described method comprises: in the client-server network, maintenance has the system of grid managers, described grid managers has hierarchical relationship, the relation of described each grid managers is stored in each described system with grid managers, wherein, each described relation all is classified as higher level or subordinate, wherein, each grid managers safeguard with the tabulation of all superior relation of other grid managers and with the tabulation of all inferior relation of other grid managers, and, relation between the network manager can be modified, so that network manager has new superior relation or inferior relation.

On the other hand, the present invention is a feature with a kind of system with grid managers, this system comprises: the network of computer system, each described computer system comprises grid managers, each described grid managers and other grid managers have hierarchical relationship, the relation of described each grid managers is stored in each described computer system, wherein, each described relation all is classified as higher level or subordinate, wherein, each grid managers safeguard with the tabulation of all superior relation of other grid managers and with the tabulation of all inferior relation of other grid managers, and, relation between the network manager can be modified, so that network manager has new superior relation or inferior relation.

On the other hand, to handle service method in network system be feature with a kind of in the present invention, described method comprises: in network, start the execution of the service of first on first computing machine, described first service uses described network to carry out the processing that the location that one or more computational resource carried out, reservation, distribution, monitoring and the releasing that is used for one or more application distributed at least.Described method also comprises: read in order to notify the file of described first service with the relation of second service by described first service, wherein, described first service has inferior relation with described second service, and described inferior relation means described second service can serve the inquiry that sends for available computational resources to described first.Described method also comprises: set up from described first and serve described second first communication channel of serving; And accept to serve opening of the described first second communication channel of serving from described second, wherein, each service safeguard with the tabulation of all superior relation of other service and with the tabulation of all inferior relation of other service, and, relation between the service can be modified, so that service has new superior relation or inferior relation.

Embodiment can comprise one or more in the following feature.Described method comprises: the message that receives the inferior relation of described first service of cancellation and described second service; Close described first and second communication channels; Receive to generate the message of serving the inferior relation that resides in the service of the 3rd the 3rd computing machine from described first; Foundation is served the described the 3rd third communication channel of serving from described first; And accept to believe opening of channel from the described the 3rd four-way of serving described first service.In some cases, set up first communication channel and also comprise: determine whether described second service makes response to the request of setting up first communication channel, and, if do not have, then after the preset time section, be established to the communication channel of described second service.

On the other hand, to handle service method in network system be feature with a kind of in the present invention, described method comprises: in network, startup resides in the execution of the service of first in first computing machine, and described first serves the described network of use carries out at least to being used for the location that one or more computational resource carried out of one or more application. the processing of reservation, distribution, monitoring and releasing distribution; Start the execution that resides in the service of second in second computing machine; Read in order to notify the file of described second service by described second service with the relation of described first service, wherein, described second service has inferior relation with described first service, wherein, described inferior relation is meant that described first service can be to the inquiry of the described second service transmission for available computational resources.Described method also comprises: set up from described second and serve described first first communication channel of serving; And set up from described first and serve the described second second communication channel of serving, wherein, each service safeguard with the tabulation of all superior relation of other service and with the tabulation of all inferior relation of other service, and, relation between the service can be modified, so that service has new superior relation or inferior relation.

Embodiment can comprise following feature.Described method also comprises: the message that is received the relation of described second service of cancellation and described first service by described second service; Close described first communication channel; Described second communication channel is not made response; Receive to create the message of serving the inferior relation of the 3rd service from described second; Foundation is served the described the 3rd third communication channel of serving from described second; And set up from the described the 3rd four-way of serving described second service and believe channel.

On the other hand, the present invention is a feature with a kind of network system, described system comprises: two or more computing machines, its each all be configured to operation service, the processing that the location that one or more computational resource carried out that is used for one or more application, reservation, distribution, monitoring and releasing are distributed is carried out in described service at least.Described system also comprises: the network of described service, described network is configured so that have a superior relation from first service of described service and second service from described service, and, described second service has inferior relation with described first service, wherein, described first service is configured to by waiting for for the state of checking described second service the described network from described first response of serving the inquiry of described second service, wherein, each service safeguard with the tabulation of all superior relation of other service and with the tabulation of all inferior relation of other service, and, relation between the service can be modified, so that service has new superior relation or inferior relation.

Embodiment can comprise following feature.Described relation comprises from described first serves the first communication channel of described second service and serves the described first second communication channel of serving from described second.Described first service also is configured to locate the one or more computational resources that are used for described one or more application by sending to described second service for the inquiry of available computational resources.Described second service also be configured to remove its with the inferior relation of described first service and create new, with the 3rd superior relation of serving.

These and other embodiment can have following advantage.Be used for the Dynamic Tree Structure of mesh-managing by use, can obtain quick and healthy and strong grid computing environment.

To set forth one or more embodiments of the detail of the present invention in accompanying drawing below and the explanation.Other features, objects and advantages of the present invention will become clear by instructions and accompanying drawing and by claims.

Description of drawings

Fig. 1 is the block scheme of grid computing environment.

Fig. 2 is the process flow diagram of resource that is used for finding and be reserved in the grid computing environment of Fig. 1.

Fig. 3 is used for installing, move and remove the process flow diagram of using at the grid computing environment of Fig. 1.

Fig. 4 is the block scheme of the computer equipment in the grid computing environment of Fig. 1.

Fig. 4 A is the process flow diagram of application that is used for being enabled in the computer equipment of Fig. 4.

Fig. 5 is the process flow diagram of grid managers that is used for being enabled in the grid computing environment of Fig. 1.

Fig. 5 A is the block scheme of grid computing environment that has increased Fig. 1 of another computer equipment.

Fig. 6 is the block scheme that is used for exemplary grid pattern user interface (GUI) assembly of visual grid computing environment.

Fig. 7 is the block scheme of grid browser component.

Reference numeral similar in different accompanying drawings is indicated similar unit.

Embodiment

As shown in Figure 1, the computational resource that Service Management in the grid computing environment 100 is used to use. grid computing environment 100 can be distributed separately with the calculating of execution application or a distribution type computational resource of data retrieval task. and computational resource comprises

computer equipment

12,14,16,18,20 use network 8 communicate by letter with 22. computer equipments. and application has scalable calculation requirement. for example, use the computer equipment 12 in the

grid computing environment

100,14,16,18,20 and 22 example application is an internet price configurator (configurator). and computer equipment 12 provides access to netwoks to pricing information via the web browser on the computer equipment that is connected to the internet to the user. and this web browser can be can displaying contents and/or carry out such as webpage, any application of the application of media file and program is such as Netscape

Microsoft Internet

And similar application.

In this example, the webserver on the computer equipment 12 provides the calculating parameter of each price that pricing information .IPC scheduler (dispatcher) 116 will calculate to send to respectively at computer equipment 12 to the user, 14, the IPC server of carrying out on 16 and 18 120,122,124 and 126. because the webserver and the application flexibility on the internet, user's number can change. and this has just produced the dynamic calculation requirement .IPC manager 118 of internet price configurator and the communication for service in the grid computing environment 100, so that service can according to the internet fix a price that the dynamic calculation of configurator requires and the allocation and deallocation computational resource (for example,

computer equipment

12,14,16,18,20, processor in 22). the allocation and deallocation computational resource makes computer equipment 12 by this way, 14,16,18,20 and 22 are designated as general-purpose computational resources, and can just not be exclusively used in and handle the peak demand .

IPC manager

118 and 116 collaborative works of IPC scheduler that internet price configurator is used, so that IPC scheduler 116 can be visited the resource in (have access to) network 8.

The ability of the resource in this allocation and deallocation grid computing environment 100 make IPC manager 118 can be in " as required " in case the basis on location and use available computational resources. resource is positioned, IPC manager 118 just can use service in the grid computing environment 100 with IPC server 120,122,124 and the 126 fully last .IPC schedulers 116 of computing machine that are installed in the grid computing environment 100 as application use by people such as Tuecke at open grid service infrastructure (Open GridServices Infrastructure, OGSI) Web Service Definition Language that defines in 1.0 versions (WebService Definition Language, WSDL) IPC scheduler 116 and IPC server 120 are managed and exchanged to interface, 122, information flow between 124 and 126. for example, OGSI WSDL interface can be used to the calculating parameter that calculates that is used to fix a price is sent to IPC server 120 from IPC scheduler 116,122,124 and the 126.OGSI WSDL interface result that can also be used to finish from IPC server 120,122,124 and 126 send back IPC scheduler 116. is incorporated herein by reference .OGSI WSDL interface with OGSI 1.0 versions and makes it possible to grid computing environment 100 and carry out controlled such as the application of internet price configurator, have (fault-resilient) of error recovery capabilities and the management of safety.

Use IPC server 120 at IPC scheduler 116,122,124 and 126 when carrying out users' calculating, service monitoring operation IPC server 120 in the grid computing environment 100,122,124 and 126, the utilization of resources of the computer equipment in the grid computing environment 100. service also utilizes this information to send to IPC manager 118. according to the comparison between utilization requirement and current resource load, IPC manager 118 can dynamically notify the service in the grid computing environment 100 so that be IPC server 120,122,124 and 126 distribute more resources, perhaps remove Resources allocation, so that make the utilization of resources in the grid computing environment 100 remain on the level of expectation.

Grid managers 152,154,156,160,162 and 164 reside in computer equipment 12 respectively, 14,16,18, in 20 and 22. in grid computing environment 100, grid managers is to having direction relations (directional relations), this direction relations is categorized as a grid managers higher level (superior) of another grid managers. and grid managers can have superior relation more than one with other grid managers. for example, grid managers 152 has superior relation with

grid managers

154 and 156. and grid managers can also have inferior relation (inferiorrelations) more than one with other grid managers. by these hierarchical relationships, IPC manager 118 does not need the tabulation of all computer equipments in the accesses network 8, so that use the computational resource in the grid computing environment 100. only require IPC manager 118 can visit the network address of the computer equipment (for example moving the computer equipment 12 of grid managers 152) of an operation grid managers, and this grid managers utilize it and the relation that operates in other grid managers on other computer equipment to provide dereference to other computer equipment in the grid computing environment 100 to IPC scheduler 116.

Grid managers (for example 152,154,156,160,162 and 164) safeguard with first tabulation of all superior relation of other grid managers and with second tabulation of all inferior relation of other grid managers. each grid managers is all at transmission control protocol (TCP), the communication channel of all grid managers during for example above-mentioned OGSI WSDL interface maintenance of use is tabulated to these via network 8 on HTTP(Hypertext Transport Protocol) and the Simple Object Access Protocol (SOAP) " opening forever ".These tabulations and corresponding communication channel can be modified, to allow dynamically to reconfigure trellis stage during working time. this also allows dynamically to replace the fault grid managers in level. for example, with reference to Fig. 1, if grid managers 154 breaks down, then grid managers 152 is lost it to grid managers 160 and 162 be connected. in this case, relation between the grid managers can be modified, it is new so that grid managers 152 has, superior relation with grid managers 160 and 162. similarly, grid managers 160 and 162 has new, inferior relation with grid managers 152.

As shown in Figure 2, design application start and handled 200, so that (for example using, internet price configurator) (for example at computer equipment, 12,14,16,18,20 or 22) go up before the execution, the essential resource that acquisition distributes in network 8. handle 200 and also guarantee: if similar application is attempted to start simultaneously on the same asset on the computer equipment, then these two or more application can not conflict or the phase mutual interference mutually. for example, for guarantee service quality (QoS), IPC manager 118 can require: IPC server (for example 120) is unique application of carrying out on the processor in computer equipment 14. in this case, if other is used on the processor of attempting simultaneously in computer equipment 14 and carries out, then Another Application will be disturbed.

Handle 200 and comprise that IPC manager 118 (or some other application) sends (202) requirement to computational resource, with inquiry grid managers (for example 154) thus determine in grid computing environment 100, whether there are the available resources that require coupling with these. the information about the resource in the computer equipment has been specified in these requirements, all numbers of processor as requested, desired those processors, primary memory and network speed utilize number percent. inquiry can also comprise that relevant this inquiry should be transmitted to which rank of the information of (in grid computing environment 100) level. handle 200 and comprise that grid managers 154 receives (204) these requirements.

For respond from IPC manager 118, to the inquiry of available resources, to handle 200 and comprise that grid managers 154 will require to mate (206) with the known resource of grid managers 154. these resources comprise by the direct resource (for example, processor 40) in the computer equipments 14 of management of grid managers 154. current resources available and that meet the demands, that directly managed by grid managers 154 are added in the resource-query list by grid managers 154 maintenances.

Grid managers 154 also sends these inquiries to the grid managers 160 and 162 that has an inferior relation with grid managers 154. and handle 200 and comprise that grid managers 160 and 162 satisfies the requirement of being asked and grid managers 160 and 162 that can use and known resources (for example computer equipment 18 by sending to grid managers 154 respectively, processor on 20) tabulation responds (208) if this inquiry. and these resource-query list of grid managers 160 and 162 known resources can also comprise by having the resource of the grid managers (not shown) management of inferior relation with grid managers 160 and 162. and grid managers 154 will add to from these resource-query list of the available resources of grid managers 160 and 162 in its resource-query list of the available resources that satisfy the requirement of being asked. and handle 200 and determine that (for example there is at least one resource in (210) in this resource-query list, processor 40), then grid managers 154 with this resource-query list send (214) to IPC manager 118. otherwise, (for example determine (212) grid managers 154 and higher level's grid managers if handle 200, grid managers 152) has relation, then grid managers 154 will send to (202) grid managers 152. to the inquiry of available resources in response to this inquiry, and grid managers 152 can not send it back unnecessary inquiry the grid managers 154. that has inferior relation with grid managers 152

Handling 200 addresses that comprise its corresponding grid managers in that grid managers 154 requires the tabulation of available resources and coupling, the network 8 sends (214) .IPC manager 118 together and selects resource (for example from this tabulation, resource on the computer equipment 16), and to grid managers 154 request (216) of the resource on the Management Calculation machine equipment 16 if reserve resource on the computer equipment 16. the resource on the computer equipment 16 still can be used for reserving (218) and reserving successfully, and then grid managers 154 is to IPC manager 118 transmission (220) reserved number.This reservation means and guarantees that IPC manager 118 has and distribute resource on the computer equipment 16 that asked, in the grid computing environment 100 to it. grid managers 154 use the independent processing thread of carrying out to handle to come from such as the application of IPC manager 118, to the inquiry of available resources. like this, grid managers 154 uses signal lamps (semaphore) to guarantee that identical resource (for example, processor 40) can not ask identical resource simultaneously and be assigned with a plurality of reserved number owing to different application.

If the resource in the computer equipment 16 that grid managers is determined to be asked can not be used for reserving and reserving failure, then next available resources in 118 selective listings of IPC manager and request (216) are if reserve this next one available resources. and IPC manager 118 receives herd number and does not have expired (222) from sending the time-out time (timeout) that this herd number begins to measure, the resource of IPC manager 118 in computer equipment 16 then, start (224) IPC server 122. on the processor 40 by transmitting reserved number to grid managers 156 and application file begins to start IPC server 122, and grid managers 156 reads this application file so that install and carry out IPC server 122. on computer equipment 16 subsequently

As shown in Figure 3, handling 250 computer equipments (for example 14) in grid computing environment 100 goes up to install and (for example uses, IPC server 122) so that sets up available resources for this application, use this available resources, and removing when no longer needing this resource or unloading this application so that discharge described resource is that later application is used. handle 250 and comprise that IPC manager 118 transmits (252) and comprises the code of IPC server 122 and for how to install, customization, follow the tracks of this application and remove the application file of the indication of this application, so that grid managers 154 can make computer equipment 14 turn back to virgin state after carrying out this application from computer equipment 14.

IPC manager 118 for example uses file transfer protocol (FTP) (FTP), HTTP(Hypertext Transport Protocol) or from network attached storage (Network Attached Storage, NAS) file copy sends application file to computer equipment 14. information relevant for installation and customized application IPC server 122 in this zip file as the single file such as the zip file that compresses. and this information is by little executable program or extending mark language (XML) document, and---it is by the installation in the grid managers 154 and customize extraction of engine (not shown) and translation (254)---represented. and processing 250 comprises also operation (258) this application of grid managers 154 installations (256). in installation (256), in the process that customization and execution (258) are used, institute to computer equipment 14 changes and all is logged, so that answer the request of IPC manager 118 and when stopping (260) or unloading and using at grid managers 154, grid managers 154 can be removed this application from computer equipment 14, and remove any other change that (262) are made computer equipment 14 simultaneously when this application of installation and operation. like this, computer equipment 14 reverts to it in the virgin state of carrying out before using, and all resources on the computer equipment 14 can be used by later application once more. this allows to need not to restart computer equipment 14, resource just becomes available after operation is used. and these changes comprise that branchs is used in storage and moves storer (for example 32) space and other the change of application code, for example to the distribution of communication port.

In some instances, can move a plurality of application simultaneously on the resource in single computer equipment (for example 14). according to their resource requirement, the application of grid computing environment 100 is partly classified. and to use and some of computer equipment are only changed to carrying out required the first time of the application of its classification in order to move, and execution does not subsequently need these changes. in these examples, grid managers 154 only makes a change for carrying out for the first time. in addition, when unloading was used, grid managers 154 was only removed the change for the last application that is performed and stops.

Install after the application on the computer equipment in grid computing environment 100, grid managers is configured to ask and starts or stops the processing of these application. in internet price configurator (IPC) examples of applications, grid managers 154 is configured to start or stop IPC server 122.IPC manager 118 utilizations according to the resource in the current grid computing environment 100 on the computer equipment 14 after on computer equipment 14 IPC server 122 being installed, the request grid managers starts or stops the IPC server in the grid computing environment 100. after the IPC server 122 on having stopped computer equipment 14, IPC manager 118 is waited for preassigned time quantum, if indicating, the current then utilization of resources need not start IPC server 122 once more, then IPC manager 118 request grid managers 154 unloading IPC servers 122. in addition, as previously mentioned, grid managers monitor such as computer equipment 14, (for example the IPC server 120 in the operation application, 122,124 and 126) utilization of resources on the computer equipment, and utilize information to send to IPC manager 118. this

In many examples, to the application on the resource in the computer equipment handle be controlled to be operating system (OS) peculiar. grid computing environment 100 is configured to the different operating system on the process computer equipment. in addition, grid computing environment 100 is designed to (for example handle different application, internet price configurator), it needn't be redesigned in order to carry out in grid computing environment 100. and grid managers is used with universal mode control and is handled, this universal mode has reduced the interdependent property between exploitation grid managers code and the Application and Development code. provide interface to application code, so that grid managers can be found, control (for example starts, stop, suspend, continue) and check or monitor the state of handling of using. handling for the operating system that is subjected to the domination of operating system or hosted environment provides interface, and this interface comprises three aspects. an aspect of interface is a deal with data, such as handling sign, state, resource consumption is (such as CPU (central processing unit) (CPU), storer, other resource that socket binding (socket binding) or application can be used) degree, and by the application specific data of process data scheme definition.

The second aspect of interface is a bookkeeping, such as other operation that starts, stops, waiting for, continues, changes priority and defined by the bookkeeping of being supported.

The third aspect of interface is control binding and definition, such as process data scheme, the bookkeeping of being supported and communication bindings. because the application of not all operation in grid computing environment 100 can be visited this identical information and ability aspect three, therefore using to grid managers provides each to use the inquiry supported and the tabulation of order.

The interface that offers application code is that application programming interfaces (API) .API is the group of methods (being embedded in the software code) by the grid managers software dictates, utilize these methods, the programmer who writes application program (for example internet price configurator) can handle the request from grid managers.

As shown in Figure 4, IPC server 122 comprises that API 302 and document 304. are because API 302 is applicable to dissimilar application, therefore document 304 is described what request that how grid managers 154 communicate by letter with IPC server 122 and IPC server 122 is supported by API 302. and grid managers 154 read document 304. in some instances before startup IPC server 122, document 304 usefulness XML are write as, and comprise Doctype description (Document Type Description, DTD) 306.DTD is that the specifically defined .DTD that follows standard generalized markup language (SGML) rule is a kind of standard (specification), it is followed document and has identified what mark and separates paragraph, what mark sign topic headings, and how each mark will be handled. by in document 304, comprising DTD 306, grid managers 154 with DTD " reader " (or " SGML compiler ") just can be handled document 304, and can correctly translate many different types of, use the document 304. of a series of different marker codes and relevance

Shown in Fig. 4 A, grid managers 154 uses handles 350 application of installing such as IPC server 122. and grid managers 154 reads the DTD 306 in (352) document 304, so that the mark in the identification document 304. grid managers 154 reads the document 304 of (354) usage flag, so that identification is used for the messaging parameter of communicate by letter with IPC server 122. grid managers 154 is set up communicating by letter of (356) and IPC server 122 according to the standard of messaging parameter. and grid managers 154 use messaging parameters are communicated by letter (358) with IPC server 122, so that transmission is such as " startup ", " stop " and " you? " free time request.

Can on network 8, (for example carry out any application, internet price configurator) before, grid managers 152,154,156,160,162 and 164 at

computer equipment

12,14,16,18, on 20 and 22 by asynchronous starting, and the relation of they and other grid managers is established. as shown in Figure 5, handle the relation between the 400 initialization grid managers. for each grid managers (for example, grid managers 154), grid managers 154 starts on computer equipment 14 by reading (402) property file. and property file comprises the address list of the computer equipment with grid managers, if described grid managers and grid managers 154 have superior relation. this tabulation is described as first tabulation with all superior relation of other grid managers before. (404) have specified higher level's grid managers (for example grid managers 152) in this address list, then grid managers 154 requests (406) (for example are opened to higher level's grid managers, 152) if communication channel. grid managers 152 has been activated, then grid managers 152 is by accepting to make response from the request of opening communication channel of grid managers 152. and handle 400 and comprise that grid managers 154 detects (408) and has the grid managers of inferior relation with grid managers 154 (for example grid managers 160 from being identified as, 162) if any communication channel request. handle 400 and determine that there are some requests in (410), then grid managers 154 allows to come from inferior grid managers (for example 160,162) communication channel. handle 400 and comprise that grid managers 154 checks whether (414) exist any hang-up, for request to the communication of grid managers with superior relation. if having the request of any hang-up, then grid managers 154 requests (406) are to communication channel of grid managers. and resource-query (as previously mentioned) and " heartbeat " message between the grid managers that these communication channels are used between grid managers are all being worked to guarantee each grid managers in the grid computing environment 100.

In case have the grid managers 152,154,156,160,162 and 164 operations of the relation of having set up, grid managers just is used to the proper handling of grid computing environment 100. the function of grid managers usually is enhanced in the life cycle of grid computing environment 100. and it is impossible or inconvenient often to close grid computing environment 100 and start the grid computing environment 100 with enhancement function.Configuration grid managers 152,154,156,160,162 and 164 so that only can cause minimum influence when changing to the user of grid computing environment 100. in order to realize this transparency, for user interface provides API, so that the keeper of grid computing environment 100 can be individually or is all visited each grid managers 152 together, 154,156,160,162 and 164. these API are static, reason is that it only comprises a kind of method, promptly, the character string .API that comprises the order of being keyed in by the keeper is dynamic, and reason is that this character string can comprise many different orders.

In some cases, grid managers is to use the exploitation of Java programming language. in these situations, can be distributed to the newer command of grid managers by dynamically load java class support new or that revise via Classloader (classloader). need not to close grid managers in the grid computing environment 100 and just can realize this dynamic access code. use java class loader, during the example of the class of each generating mesh manager, the definition of class and behavior can both be updated, thereby provide new function to grid computing environment 100.

The another kind of mode that need not to close grid computing environment 100 and dynamically revise the function of grid computing environment 100 is to change hierarchical relationship between the grid managers, remove grid managers or add new grid managers. the API that provides for the management of grid computing environment 100 also is configured to send to each grid managers and has the character string of order to delete existing relation or to add new relation.

Keeper for grid computing environment 100, grid managers and application on the computer equipment in the visual grid computing environment 100, and be useful to have the grid managers on other computer equipment of a part of form operation mesh-managing level of grid managers of one or more levels inferior relation and to use with this grid managers. the view of these computer equipments is called as grid landscape (grid landscape). as shown in Figure 6, the grid pattern user interface (GUI) 500 that is used for visual grid landscape such as grid computing environment 100 comprises that the GUI unit .GUI 500 of the tissue of the visual service that moves provides the fenestral fabric with row and row on computer equipment. the line display service, with respect to the application under the service, the type of service and specific Service Instance and the described service of laminated tissue successively. in each service instance row and the grid computing environment 100, represent that its position by the place of instantiation is associated. in the case, the computer equipment in the grid landscape is shown in tabulation. specifically, GUI 500 has three computer equipments 12 of expression, 14 and 16 three row .GUI 500 illustrate grid managers 152 and moves on computer equipment 12, and

inferior grid managers

154 and 156 also illustrates internet price configurator services at operation .GUI 500 on

computer equipment

14 and 16 respectively and moves on computer equipment 12. and these internets configurator services of fixing a price comprise IPC scheduler 116, IPC server 120 and IPC manager 118.

The feedback that utilization comes from grid managers and internet price configurator (or other is used) service dynamically refreshes GUI 500, so that present new service to the keeper in GUI 500. similarly, from GUI 500, remove pent service.

As shown in Figure 7, grid browser component 600 is the composite figure user interfaces (GUI) that are used for browsing the grid managers on the computer equipment of grid computing environment 100. assembly 600 shows the figure with bent limit and summit. and the computer equipment in the vertex representation grid computing environment 100, and the directional correlation of the grid managers on two computer equipments (summit) in the grid computing environment 100 is represented on bent limit. this association be hierarchical (promptly, superior/inferior). each summit shows the network address and the current application that moves of computer equipment on this computer equipment. for example, assembly 600 illustrates has IPC server 118,120,122 and 124

computer equipment

12,14,16,18,20 and 22. in another example (not shown), grid browser component 600 illustrates non-level between the grid managers, the association of equity (peer-to-peer), and use nondirectional limit to represent these associations.

Grid browser component 600 is context-sensitive (context sensitive). depends on the relation (for example superior/inferior) between the grid managers on the computer equipment, considers user's browsing histories and travel through computer equipment.

By clicking the summit of representing computer equipment (for example computer equipment 14) among the GUI 600, grid managers and application that the user can use GUI 500 automatically to check to move on this computer equipment, and the grid managers that has inferior relation with this grid managers. the user can select computer equipment, and sees its grid managers and the relation between other grid managers. and use the software that generates GUI500 and 600 to realize this connection between GUI 500 and 600.

Network 8 can be realized in many ways. network 8 comprises the network of any kind of, such as the Internet, Local Area Network or other local network, private network, public network, plain-old telephone system (POTS) or other similar wired or wireless network, and any combination. can use such as encryption, the mechanism of security protocol or the similar mechanism of other type guarantee the safety by the communication of network 8. can comprise the communication linkage of any kind of by the communication of network 8, such as modem links, the Ethernet link, cable, the point-to-point link, infrared connection, fiber optic links, wireless link, cellular link;

, satellite link and other similar link, and any combination.

For the ease of explaining, network 8 to be simplified. network 8 can comprise more or extra cell still less, for example network, communication linkage, acting server, fire wall or other security mechanism, Internet Service Provider (ISP), gatekeeper, gateway, switch, router, hub, client terminal and other unit.

Computer equipment

12,14,16,18,20 and 22 uses a kind of in the multiple different networking protocol to communicate on medium 10. and for example, a kind of agreement is the TCP (TCP/IP) that combines with SOAP (Simple Object Access Protocol).

Embodiments of the invention can be realized by enough Fundamental Digital Circuits, perhaps use computer hardware, software, firmware or its make up to be realized. embodiments of the invention can be implemented as computer program, promptly,---for example with node readable storage device or use transmitting signal---the visibly specific computer program of using information carrier, it is by for example programmable processor, the data processing equipment of computing machine or many computing machines is carried out, or control the operation of described data processing equipment. computer program can enoughly comprise compiling or interpretive language, any type of programming language is write as, and can be to comprise as stand-alone program or as module, assembly, any form configuration computer program of the unit that subroutine or other are adapted at using in the computing environment. computer program can be configured to carry out on a computing machine, perhaps on many computing machines of the three unities, carry out, perhaps be distributed in a plurality of places and carrying out on interconnective many computing machines by communication network.

The method step of the embodiment of the invention can be carried out by the programmable processor of one or more computer program, so that carry out function of the present invention by output is operated and generated to the input data. method step can also be carried out by the dedicated logic circuit of for example FPGA (field programmable gate array) or ASIC (special IC), and equipment of the present invention also can be implemented as described dedicated logic circuit.

For instance, the processor that is suitable for computer program comprises general and special microprocessor, and any one or a plurality of processor of the digital machine of any kind of. usually, processor will receive instruction and data from ROM (read-only memory) or random access memory or both. and the essential elements of computing machine is processor that is used to execute instruction and the one or more memory devices that are used for storage instruction and data. usually, computing machine also will comprise one or more mass-memory units that are used to store data, perhaps computing machine is operably connected to this mass-memory unit so that receive data or transmit data to it from it, perhaps both; Mass-memory unit is for example disk, magneto-optic disk or CD.Information carrier that be fit to specialize computer program instructions and data comprises the nonvolatile memory of form of ownership, for instance, comprises for example semiconductor memory apparatus of EPROM, EEPROM and flash memory device; The disk of internal hard drive or portable hard drive for example; Magneto-optic disk; And CD-ROM and DVD-ROM dish. processor and storer can be replenished by dedicated logic circuit, perhaps with incorporated in special purpose logic circuitry.

For mutual with the user is provided, embodiments of the invention can have CRT (cathode-ray tube (CRT)) for example or LCD (LCD) monitor, be used for can the keyboard of input is provided and for example implement on the computing machine of the indicating equipment of mouse or tracking ball to computing machine to the display apparatus of user's display message with by its user. also can use the equipment of other kind that mutual with the user is provided; For example, the feedback that provides to the user can be any type of sensory feedback, for example, and visual feedback, audio feedback or tactile feedback; And, can be received in any form from user's input, comprise the sense of hearing, voice or sense of touch input.

Embodiments of the invention can comprise aft-end assembly, or comprise middleware component, or comprise front end assemblies, or comprise this rear end, realize in the computing system of the combination in any of middleware or front end assemblies, wherein, aft-end assembly is for example as data server, middleware component is an application server for example, front end assemblies is for example to have the client computer that the user can carry out interactive graphical user interface or web browser by the realization of itself and the embodiment of the invention. what the assembly of system can be by for example communication network, any type of digital data communication or its medium interconnect. and the example of communication network comprises LAN (Local Area Network) (" LAN ") and wide area network (" WAN "), for example the Internet.

Computing system can comprise client and server. and client and server and is generally undertaken by communication network usually away from each other alternately. and the relation of client and server relies on the computer program that moves on each computing machine and have the client-server relation each other and produces.

A plurality of embodiment of the present invention has been described. still, should be appreciated that, under the situation that does not break away from the spirit and scope of the present invention, can carry out various modifications. other embodiment is within the scope of the appended claims.

Claims

1. a maintenance has the method for the system of grid managers, comprising:

In the client-server network, safeguard to have the system of grid managers, described grid managers has hierarchical relationship, the relation of described each grid managers is stored in each described system with grid managers, wherein, each described relation all is classified as higher level or subordinate

Wherein, each grid managers safeguard with the tabulation of all superior relation of other grid managers and with the tabulation of all inferior relation of other grid managers, and the relation between the network manager can be modified, so that network manager has new superior relation or inferior relation.

2. system with grid managers comprises:

The network of computer system, each described computer system comprises grid managers, each described grid managers and other grid managers have hierarchical relationship, the relation of described each grid managers is stored in each described computer system, wherein, each described relation all is classified as higher level or subordinate

3. in network system, handle service method for one kind, comprising:

In network, start the execution of the service of first on first computing machine, described first service uses described network to carry out the processing that the location that one or more computational resource carried out, reservation, distribution, monitoring and the releasing that is used for one or more application distributed at least;

Read file by described first service, so that notify described first service with described first service and the relation of second service, wherein, described first service has inferior relation with described second service, and described inferior relation means described second service can serve the inquiry that sends for available computational resources to described first;

Foundation is served described second first communication channel of serving from described first; And

Acceptance is from described second the opening of second communication channel of serving described first service,

Wherein, each service safeguard with the tabulation of all superior relation of other service and with the tabulation of all inferior relation of other service, and the relation between the service can be modified, so that service has new superior relation or inferior relation.

4. method as claimed in claim 3 also comprises:

Receive the message of the inferior relation of described first service of cancellation and described second service;

Close described first and second communication channels;

Receive to generate the message of serving the inferior relation that resides in the service of the 3rd the 3rd computing machine from described first;

Foundation is served the described the 3rd third communication channel of serving from described first; And

Acceptance is believed opening of channel from the described the 3rd four-way of serving described first service.

5. as claim 3 or 4 described methods, wherein, set up first communication channel and also comprise: determine whether described second service makes response to the request of setting up first communication channel, and, if no, then after the preset time section, be established to the communication channel of described second service.

6. in network system, handle service method for one kind, comprising:

In network, start the execution that resides in the service of first in first computing machine, described first service uses described network to carry out the processing that the location that one or more computational resource carried out, reservation, distribution, monitoring and the releasing that is used for one or more application distributed at least;

Start the execution that resides in the service of second in second computing machine;

Read file by described second service, so that notify described second service with described second service and the relation of described first service, wherein, described second service has inferior relation with described first service, wherein, described inferior relation is meant that described first service can be to the inquiry of the described second service transmission for available computational resources;

Foundation is served described first first communication channel of serving from described second; And

Foundation from described first serve described second the service the second communication channel,

7. method as claimed in claim 6 also comprises:

Receive the message of the relation of described second service of cancellation and described first service by described second service;

Close described first communication channel;

Described second communication channel is not made response;

Receive to create the message of serving the inferior relation of the 3rd service from described second;

Foundation is served the described the 3rd third communication channel of serving from described second; And

Foundation is believed channel from the described the 3rd four-way of serving described second service.

8. network system comprises:

Two or more computing machines, its each all be configured to operation service, the processing that the location that one or more computational resource carried out that is used for one or more application, reservation, distribution, monitoring and releasing are distributed is carried out in described service at least;

The network of described service, described network is configured so that have a superior relation from first service of described service and second service from described service, and, described second service has inferior relation with described first service, wherein, described first service is configured to by waiting for for the state of checking described second service the described network from described first response of serving the inquiry of described second service

9. system as claimed in claim 8, wherein, described relation comprises from described first serves the first communication channel of described second service and serves the described first second communication channel of serving from described second.

10. system as claimed in claim 8 or 9, wherein, described first service also is configured to locate the one or more computational resources that are used for described one or more application by sending to described second service for the inquiry of available computational resources.

11. system as claimed in claim 8 or 9, wherein, described second service also be configured to remove its with the inferior relation of described first service and create new, with the 3rd superior relation of serving.