CN106462444A - Architecture and method for cloud provider selection and projection - Google Patents

Abstract

The invention relates to an architecture and method for cloud provider selection and projection. According to the invention, an architecture and method are provided for selecting and projecting a data center. For example, one embodiment of an apparatus comprises: a broker component to manage data center information, the broker component to receive updates containing the data center information and to respond to queries to provide the data center information; a selection engine to perform data center selections and/or recommendations for a cloud user based on data center information read from the broker component, wherein the data center selections and/or recommendations are made based on data center requirements and/or preferences specified by the end user; and a projection component to interface with a selected data center and transmit data to configure the data center on behalf of the user upon selection of the data center from the selections and/or recommendations made by the selection engine.

Description

The framework of the selection of cloud provider and projection and method

Background technology

Cross-Reference to Related Applications

This application claims the rights and interests of following patent application：The U.S. Patent application No.14/ that on January 2nd, 2015 submits to 588,613；U.S. Provisional Patent Application No.61/924,122 that on January 6th, 2014 submits to；U.S. that on January 2nd, 2015 submits to State patent application No.14/588,620；U.S. Provisional Patent Application No.61/924,125 that on January 6th, 2014 submits to；2015 The U.S. Patent application No.14/588,626 that on January 2, in submits to；The U.S. Provisional Patent Application that on January 6th, 2014 submits to No.61/924,128；The U.S. Patent application No.14/588,631 that on January 2nd, 2015 submits to；On January 6th, 2014 submits to U.S. Provisional Patent Application No.61/924,133；The U.S. Patent application No.14/588,658 that on January 2nd, 2015 submits to； U.S. Provisional Patent Application No.61/924,137 that on January 6th, 2014 submits to；The United States Patent (USP) Shen that on January 2nd, 2015 submits to Please No.14/588,607；And U.S. Provisional Patent Application No.61/924 that on January 6th, 2014 submits to, 143, these applications All it is hereby incorporated herein by accordingly.

Technical field

Present invention relates generally to the field of data handling system.More particularly, it relates to for the choosing of cloud provider The system and method selected and project.

Association area explanation

Cloud computing can service (IaaS) using infrastructure, platform services (PaaS) and software services (SaaS) Model is providing.Any one in these models can be realized in " data center " based on cloud, described based on cloud " in data The heart " is made up of various computing resources (for example, server, router, load equalizer, switch etc.).

IaaS is most basic model.IaaS provider provides physical computer or virtual machine (that is, using virtual Machine) and other resources such as virtual machine disks image library, storage resource (including the storage based on file), fire wall, load balancing Device, IP address, VLAN (VLAN) and software kit.IaaS provider can be dynamic larger pond in the data center from installing These resources are supplied on state ground.In order to dispose its application program, operating system image and application software are arranged on by cloud user In cloud resource.In this model, cloud user attended operation system and application software.Generally, cloud provider is based on and is divided Join and the stock number that consumed and to user's charging.

In PaaS model, complete calculating platform is paid by cloud provider, this calculating platform generally include operating system, Web server, programming language performing environment database.Application developer is developed and runs software in this cloud platform Solution, and there is not the cost related to purchase and fundamentals of management hardware and software layer and complexity.In some PaaS tool During body is implemented, basic resource (for example, calculate, storage etc.) automatically scale to mate application requirement so that cloud user no Need manually Resources allocation.

In SaaS model, cloud provider installs and maintenance applications software in cloud, and cloud user is from cloud client Access software (sometimes referred to as " software on demand " model).This eliminates installation and operation application on the cloud user computer of oneself The demand of program, and then simplify maintenance and support.Compared with the application journey locally executing, cloud application program provides actually no The extensibility limiting, this to be realized on multiple virtual machines by being cloned into task during runtime, thus meeting change The work requirements changed.Load equalizer by work transparently to users on virtual robot arm distribution (user only sees single access Point).

Brief description

The present invention can be more fully understood in conjunction with drawings below from detailed description below, wherein：

Figure 1A shows cloud analysis and an embodiment of projection service；

Figure 1B illustrates the details related to cloud provider market；

Fig. 2A shows system architecture according to an embodiment of the invention；

Fig. 2 B to Fig. 2 C shows method according to an embodiment of the invention；

Fig. 3 shows the legend of data center's arbitration (arbitrage) employed in one embodiment of the present of invention；

Fig. 4 shows an embodiment selecting exchange architecture；

Fig. 5 A to Fig. 5 D shows related to the virtualization including graphic user interface and one embodiment of projecting subassembly In addition details；

Fig. 6 shows many employed in one embodiment of typical data center for projecting to virtual data center Individual logical layer；

Fig. 7 shows the other details related to an embodiment of overall intermediary；

Fig. 8 shows that virtual data center covers an embodiment of (overlay)；

Fig. 9 to Figure 10 shows an embodiment of the distributed file system engine for migrating data center；

Figure 11 A to Figure 11 B shows an embodiment of the shadow storage system for migrating data center；

Figure 12 A to Figure 12 C shows the gateway for migrating data center and network in one embodiment of the present of invention Infrastructure；

Figure 13 A to Figure 13 B shows that data of acting on behalf of according to an embodiment of the invention collects process；

Figure 14 shows overall intermediary and the other details with an embodiment selecting engine communication.

Specific embodiment

Described below is the embodiment of the equipment, method and machine readable media selecting for cloud service and projecting.Whole In individual description, for illustrative purposes, set forth many specific detail herein thoroughly to understand the present invention.However, for this Field it is obvious to the skilled person that the present invention can be put into practice in the case of some during there are no these specific detail. In other cases, for avoiding the ultimate principle of the fuzzy present invention, it is well known that construction and device not shown or in form of a block diagram Illustrate.

The embodiment of invention as described herein is carried using increasing cloud service for the user that those move to cloud For business.Specifically, these embodiments are included independent of the true cloud service provider providing cloud footprint, and are used for managing and move The mechanism of dynamic data center.In one embodiment, cloud footprint is IaaS footprint；However, the basic principle of the present invention also can be The data center of PaaS service or SaaS service is entirely provided to realize.

Figure 1A shows cloud analysis according to an embodiment of the invention and the high shelf of projection service (CAPS) 100 Structure.As described in detail, CAPS 100 is capable of multiple powerful models, is carried including to by cloud provider 121 to 124 For data center miscellaneous arbitration ability to create the price of optimum, performance, availability for virtual data center And/or geographic range.Specifically, an embodiment of CAPS 100 is according to the requirement specified by cloud user 111 to 115 Cost data, resource data, performance data, geographic range data, reliability data and/or any other related cloud are provided Business's variable is analyzed.Once being assessed to related data, CAPS 100 will automatically select for cloud user One or more cloud providers.Alternatively or in addition, CAPS 100 can be by executing cloud arbitration, using between cloud provider Measured difference (for example, once occurring uneven between cloud provider, including in cloud provider price, performance, service water Flat agreement or other can measure the combination occurring between tolerance that difference is the match trading of capitalization) recommending one group of " candidate " cloud Provider.Terminal use subsequently can be selected among the cloud provider candidate being recommended.

As discussed further below, an embodiment of CAPS 100 includes virtualizing and projects logic, so that data center Resource virtualizing, as soon as and once making the decision moving to another cloud provider from Ge Yun provider, realize in data The heart migrates (for example, with reference to the virtualization shown in Fig. 2A and projecting subassembly 231).Specifically, an enforcement of CAPS 100 Example generates virtualization or the logical expressions of all data center resource, and this data center resource includes but is not limited to, router, Switch, load equalizer, WAN accelerator, fire wall, vpn concentrator, DNS/DHCP server, live load/virtual machine, File system, network attached storage system, object memories and backup of memory, name just a few.Should " virtual data center " table Show the atom component reflecting including data center, and manage the basic debugging mode of each logic device.CAPS 100 with Pass through afterwards to represent virtualization and be transformed into the form realized needed for typical data center, or by directly executing in cloud provider Virtual data center (for example, using the embodiment of Full-virtualization, following article combines described in Fig. 9) is by virtual data center It is projected on new typical data center.

There are thousands of small-sized cloud service provider, they are the same with large-scale provider can to hand over to their client Pay IaaS, but these small-sized providers be considered to be scrappy type or zonal.Consider the chart shown in Figure 1B, it illustrates Currently main North America cloud service provider.Notice that the logarithm of the market share reduces.By soliciting the long afterbody on curve, CAPS 100 become " the market manipulation business " for aggregated data center service.In one embodiment, CAPS 100 is used for using intermediary's model These data center services of purchase and sale, arrange transaction between buyer and seller, and collect servant in execution transaction Gold.This is all favourable for cloud user 111 to 115 and cloud provider 121 to 124, because this achieves runs through scrappy type city The movement of field, and create accumulation type market, and need not purchase merely.

Cloud seller 121 to 124 can based on cost and its dependent variable such as persistent period (the concrete cost of finite time section), SLA, geographical position, Internet resources (being applied to some distributed application programs) and/or be exclusively used in some application journeys The hardware resource of sequence and offer to CAPS platform 100.This data can enter one by the historical record from conventional client trading Step amplification (for example, using client feedback or other rating systems).The concept of " arbitration " is therefore extended, with by bigger Cloud seller's feature list is mating the requirement of cloud buyer.One simple example is price and persistent period.Seller may be only Capacity is had to be in a certain discount rate (for example, because another client is pre- in following given time in a specific duration Determine this capacity).This to having more mobile data center framework, or only may have short term need to their data center Some buyers feasible.However, generally, come into force with the law of large numbers, it is every that CAPS 100 may have increasing probability Individual seller finds buyer, so that both making a profit.

As shown in Figure 2 A, an embodiment of CAPS 100 includes multiple assemblies, and the plurality of assembly includes overall intermediary 210th, one group of user selects engine 220 to 222 (for example, one user of each cloud user selects engine) and virtualization and projects Assembly 231.Overall intermediary 210 manages the data base 211 of all available cloud providers 121 to 124, and is provided by cloud provider Data center attribute.By way of example, data base 211 may include cost data, resource data, performance data, geography Range data, reliability data and/or any other phase related to the data center operated by cloud provider 121 to 124 Pass information.Data base 211 oneself can be dynamically updated by cloud provider 121 to 124 or by CAPS 100 member statically more Newly.For example, if particular cloud provider has changed the cost structure of its data center resource with regard to specific duration, then should Cloud provider can change the cost/persistent period updating the data in storehouse 211.Similarly, if cloud provider has upgraded it Hardware/software, or opened new data center in new position, then this cloud provider updatable data storehouse 211 is to reflect These changes.

In one embodiment, overall intermediary 210 exposes an application programming interface (API) and enables to data base 211 renewals.Cloud provider 121 to 124 subsequently can be dynamically updated data base 211 using API.This can be for example via being arranged on The CAPS software at place of cloud provider 121 to 124 each, and/or via can be by browser client in the access of cloud provider Web server realizing.Static provider updates and also can realize via API.

In one embodiment, execution user selects engine 220 to 222 to select come the data center to carry out each cloud user Select/recommend.In example shown in Fig. 2A, execution selects engine 220 to select/recommend come the data center to implement user A；Execution Engine 221 is selected to select/recommend come the data center to implement user B；And execution selects engine 222 to implement the number of user C Select/recommend according to center.In operation, each selects the data center that engine 220 to 222 is provided with corresponding cloud user to require (for example, cost, performance, reliability, geographical position etc.), and the time of subsequent those requirements of identification and matching from intermediate database Xuan Yun provider.

Following article is discussed more fully, and each selects engine 220 to 222 can generate the cloud provider 121 that coupling user requires (for example, the cloud provider positioned at list top is than positioned at list bottom for preferred list (prioritized list) to 124 These requirements are more mated in cloud provider).This list can as a Zu Yun provider " recommendation ", and be recommended to explain together with for these The comparative analysis of reason are supplied to terminal use together.Alternatively, in one embodiment, selection engine can be for user Automatically select one of cloud provider (and as discussed below, being migrated between data center).

In one embodiment, select engine 220 to 222 from overall intermediate database 211 periodically and/or automatically Ground receives to update and is selected and/or recommendation with providing new data center.For example, if a particular cloud provider 121 to 124 is aobvious Write and reduce its cost of serving, then this may result in selection engine and selects this cloud provider, and/or by this cloud provider It is placed in the top (so that it is reasonable to move to this new cloud provider) of preferred list.Select engine to receive to update and generate data The frequency of center preferred list can change according to the preference of the embodiment of CAPS 100 and/or each cloud user.

Fig. 4 shows the additional detail related to selecting engine 220, and this selection engine includes data center's order of priority and patrols Collect 420, this data center's prioritization logic is used for sending inquiry to intermediary 210, this inquiry indicates that data center requires (example As by user's request/preference 425 defined).For example, data center's prioritization logic 420 can be based on user input 425 Send inquiry, this inquiry specifies that it is only pointed in specific geographical area and has some abilities (for example, load balance, automatically Fail-over capability) data center be interested in.Therefore, needed for the candidate data center being provided by intermediary 210 is limited to have Those of parameter.

As indicated in fig. 4, data center's prioritization logic 420 subsequently (can be included based on various weighing groupwares Arbitration assembly 401, performance component 402 and reliability assembly 403) and priority is distinguished to candidate data center.Although in figure 3 Illustrate only three assemblies, but may include various other/extra assemblies in data center's prioritization logic 420, with When still comply with the present invention basic principle (for example, geographical position, terminal use data center grading etc.).

In one embodiment, weight be based on user-defined requirement/preference 425 and be assigned to each assembly 401 to 403.For example, if a particular cloud user is mainly interested in low-cost data center service, then the weight of arbitration assembly 401 Bigger than performance composition 402 and availability assembly 403.Another cloud user is likely to can be interested in low cost, but may Performance component 402 and/or the minimum requirements of reliability composition 403 can be specified.In this case, data center's order of priority is patrolled Volumes 420 will filter out the data centers that those are unsatisfactory for minimum requirements, and be subsequently based on cost and to remaining candidate data Priority is distinguished at center.Another cloud user possible major concern data center reliability assembly 403, and therefore reliability group The weight of part 403 is bigger than arbitration assembly 401 or performance component 402.Various different/extra algorithm can be by data center Prioritization logic 420 is implemented, to generate prioritizing selection based on relative component weights or to recommend 410.

Return to Fig. 2A, once have selected new cloud provider, virtualization and projecting subassembly 231 arrive with regard to data center good at managing The migration of this new cloud provider.As described above, an embodiment of virtualization and projecting subassembly 231 maintains each cloud user institute " virtual data center " of the data center resource requiring represents, such as, router, switch, load equalizer, WAN accelerate Device, fire wall, vpn concentrator, DNS/DHCP server, live load/virtual machine, file system, network attached storage system, Object memories and backup of memory, name just a few.This " virtual data center " expression reflects former including data center Sub-component and manage the basic debugging mode (for example, user's specialized configuration of each device) of each logic device.At one In embodiment, it is whole needed for each virtual data center that virtualization and projecting subassembly 231 safeguard that its own is used for managing/migrate The data base 232 of data.Alternatively, virtualization and projecting subassembly 231 can be dependent on overall intermediate database 211 to store this number According to.

Once have selected new data center, virtual data center will be projected in newly by virtualization and projecting subassembly 231 Data center on.As described above, the projection to new data center may relate to (for example, specific hard based on cloud provider Part/software resource) or by virtual data center directly being executed on cloud provider (for example, using the enforcement of Full-virtualization Mode, as discussed below in conjunction with Fig. 9) and virtualization is represented and is transformed into the form realized needed for typical data center.Once Projection completes, and old data center can be turned off.

Fig. 2 B shows that a kind of data center's specification for specifying based on user and requirement select new data center Method an embodiment；Fig. 2 C show a kind of for the method from a data Center shift to another data center An embodiment.

It is turning initially to Fig. 2 B, in 250, the specification at user input data center.As used herein, " specification " includes number According to specific components and the framework at center, including such as router, switch, load equalizer, WAN accelerator, fire wall, VPN Concentrator, DNS/DHCP server, live load/virtual machine, file system, network attached storage system, object memories and The arrangement of backup bin.In one embodiment, virtualization and projecting subassembly 231 can provide the user graphical user circle Face (GUI), this graphic user interface be used for graphically select data center component and between interconnection (for example, referring to Fig. 8 And related text).GUI can be based on Web (for example, by providing via the Web page of browser access), or can To implement as independent application program.In one embodiment, virtualization and projecting subassembly 231 pass through inquire cloud user with The relevant series of problems of data center architecture and determine data center's specification.Alternatively or in addition, virtualize and project Assembly 231 can provide the user one group of data center's template constructing in advance, and can based on subscriber data center requirement therefrom Selected.Each template can be associated with a certain group of resource requirement, and/or has relative special parameter.

No matter user's how input data center specification, at 251, virtualization and projecting subassembly are all come using these specifications Build virtual data center to represent.As described above, in one embodiment, virtual representation includes all desired data center resources Abstract representation with topology layout (for example, the interconnection between resource).Virtual representation reflects the atom group including data center Part, and manage the basic debugging mode of each logic device.

At 252, user indicates to be various factors to be considered to candidate data center differentiation priority.As above institute State, this may relate to based on the preference/requirement of user makes weight be associated with variable, variable such as data center's cost, performance Data and/or reliability data.At 253, identify one group of candidate data center based on these specifications and requirement.For example, As discussed previously, if a particular cloud user is mainly interested in low-cost data center service, then cost variations Weight is bigger than performance variable and availability variable.Various different prioritization algorithm can be implemented, with based on group relatively Part weight and generate prioritizing selection or recommendation.

At 254, select in the heart data center from the candidate data being identified.In one embodiment, by cloud user Carry out this selection (for example, after checking candidate preference list).In another embodiment, this selection is for cloud user Automatically execute.

Select data center anyway, at 255, all virtual data center can be projected to selected physical data In in the heart.As described above, the projection to new data center may relate to (for example, the specific hardware/software based on cloud provider Resource) or by directly executing virtual data center (for example, using the embodiment of Full-virtualization, such as in cloud provider Discussed below in conjunction with Fig. 9) and virtualization is represented and is transformed into the form realized needed for typical data center.Once having projected Become, you can make this data center be placed in presence.

Fig. 2 C shows an embodiment of the method that existing data center moves to new data center.260 Place, receives and assesses data center and update, and at 261, makes to whether the migration at new data center reasonably determines (example As considered based on price, performance and/or reliability).As mentioned, overall intermediary 210 can connect from cloud provider 121 to 124 Receive and continuously dynamically update, and/or can statically update (that is, statically being updated by the member of CAPS 100).When these more When being newly stored in overall intermediary 210, each selects engine 220 to 222 to can perform its selection strategy, to determine to new data Whether the migration at center is favourable to terminal use.For example, the decision of migration can be based on current data center and/or other candidates Change (for example, cost, SLA, the data center's level in addition to cost, time of day, the property of restricted time section of data center The change of energy, availability etc.) and make.

Fig. 3 shows an enforcement of the decision-making process for selecting new data center in the way of graphical description Example.As illustrated, once detect the cost related to current data center already at marginal range outside (for example, as become Shown in this arbitration frame 371) so that it may generate migration event 370.This event can be that (for example, cloud provider can provide scheduled event With regard to giving advance notice of cost change), in this case, select engine can trigger migration event using scheduler program, with At particular point in time (for example, cost is expected to exceed the time point of target zone) mobile data center.As shown in frame 372, the overall situation Intermediary can be utilized the data relevant with all monitored data centers to feed continuously updated via projection.All monitored numbers Filtered by selecting engine to be based on screening criteria such as cost, performance, position and/or availability according to central. set, to reach one group Candidate data central projection.It is subsequently based on " event " and selects a particular data center from this group candidate, this event is such as The difference that Current projection fails to fall in marginal range, candidate projects and Current projection between increases to over threshold value and/or screening Standard changes.Final as a result, event is triggered (automatically or by cloud user manually) with migrating data center.

Return to Fig. 2 C, at 262, from this group candidate, select new data center, and at 263, by virtual data Central projection is to (for example, using various shadow casting techniques as herein described) in the new data center selecting.

As shown in Fig. 5 A to Fig. 5 D, an embodiment includes graphical user circle for creating and managing virtual data center Face and Command Line Interface.In one embodiment, the Virtual Controller employed in virtual data center is each via unique Figure is representing.Fig. 5 A shows one group of special pattern for representing Virtual Controller, including：Virtual data center figure 501st, gateway figure 502, network router figure 503, network switch figure 504, fire wall figure 505, load equalizer Figure 506, WAN accelerated graphicses 507, live load/virtual machine graphics 508, dns server figure 509, file system figure 510th, Dynamic Host Configuration Protocol server figure 511, backup of memory figure 512, Network Attached Storage figure 513, vpn concentrator figure 514, and object memories figure 515.

Set forth below is the respective brief description of the Virtual Controller representing by these graph images.Implement at one In example, provide the access to basic resource via the administration interface exposing by virtualization and projecting subassembly 221.

Virtual data center 501 is the program capturing the high-rise attribute geographical position projecting, SLA, level etc..This is For the not operation object being grouped data centers different for geographical position in top layer.The attribute of virtual data center may include position Put, SLA, data center's level, price classification.

Gateway router 502 is responsible for the public route on the Internet.Attribute includes WAN configuration, route entrance, route association View, interface monitoring, DNS characteristic, topology information/routing iinformation and affinity rule.

Network router 503 is responsible for the route between the all-ones subnet network in virtual data center.Multiple network routers Can be using different interfaces come instantiation, described interface is connected to different sub-networks, more like real router.Attribute may include Network configuration, route entrance, Routing Protocol, interface monitoring and affinity rule.

The network switch 504 embodies the concept of sub-network.Make the interface that different device is connected to each other in sub-network by this Solid modelling.In the case of the remote measurement that have collected each attachment means, the network switch can be managed entity, for identifying The service condition of data center and its cost and performance.Attribute may include network configuration, monitoring and affinity rule.

Fire wall 505 be the feature that generally provided by cloud provider but it is also possible to be by CAPS100 (directly or by App Shop concept) supplementary features being provided.Fire wall can provide a series of potential contributions, the including but not limited to network address Conversion (NAT), distributed denial of service (DDOS) protection and traffic monitoring.Attribute may include network configuration, firewall policy, prison Control strategy and affinity rule.

Load equalizer 506 is for mapping multiple identical live loads together for abducent purpose Device.Attribute may include network configuration, addressable end station, equilibrium strategy, monitoring and affinity rule.

WAN accelerator 507 is the service that can be used for making the data center interconnection on WAN.This element may include device, As provided the Riverbed of duplicate removal compression algorithm.These services can be supplied to cloud by cloud provider with virtual work its load form and use Family.Between two or more virtual data centers, can be pressed using the example (each website is using) of WAN accelerator The data through WAN is gone in contracting.Attribute may include network configuration, endpoint configurations, SLA, elemental user privilege, monitors and affine Property rule.

Live load/virtual machine 508 maintains the generic configuration of OS image.For multiple VM form such as VMDK, ISO, its Prime responsibility is to transmit these images.By maintaining these images all the time, transition process greatly simplifies.Attribute may include CPU class With quantity, memorizer, locally stored, operating system image, network configuration and affinity rule.

Dns server 509 provides both sides inside and outside of name IaaS data center for virtual data center Method.This dns server should be connected to naming service and the Siaras global dictionary/intermediary of trustship IaaS service provider.Attribute May include domain name, addressing, migration feature, monitoring and affinity rule.

File system 510 can be associated with Network Attached Storage (NAS).This document system can be shared resource, but Can have some associated privileges, that is, the privilege obtaining by addressability, or the potential privilege based on user.Literary composition The core feature of the migration component of part system includes Data Migration.In one embodiment, Virtual Controller support by data from One file system instance transfers to the ability of another file system instance.Migration can be than Virtual File System controller more Coordinate on high layer to carry out, but controller should provide at " gully (Sink) " and " source (Source) " mechanism to be shifted. As discussed more fully below, in one embodiment, using the distributed file system (example being not required for manual transfer data As Hadoop).Each example of file system is all automatically tied to existing node and downloads necessary local data.File The attribute of system may include size, network configuration, SLA, elemental user privilege, backup policy and affinity rule.

Dynamic Host Configuration Protocol server 511 allows provider of data center to define addressing scheme and ACL, and other is in logical data The control of intracardiac device.Attribute may include sub-network configuration, monitoring and affinity rule.

Backup of memory 512 is the core attribute of any high-availability applications, and can be local IaaS service The attributed feature of provider, or the possibly value-added features of CAPS.In the case of the latter, have problems in that data passes The defeated amount to outside typical data center, and relative cost.

Network Attached Storage 513 can be a kind of high-performance storage method, can be used for the 1st layer of IaaS cloud data center or For intracardiac in exclusive data.These controllers are used for managing these resources.Attribute may include LUN and size, RAID strategy, net Network configuration, SLA, elemental user privilege, backup policy and affinity rule.

Vpn concentrator 514 is Terminal Server Client by the end station for being connected to data center.VDI application program and other Basic security connects and using vpn concentrator or will be used as simple secure vpn end points.Attribute may include network configuration, fire wall Strategy, monitoring strategies and affinity rule.

IaaS cloud provider can provide and store, by object, the object storage capacity that figure 515 represents.Most preferably, object Storage virtualization controller will provide transforming function transformation function, and an object storage facility is mapped to another object storage facility.Eventually The responsibility of end application program is probably to process API using the abstract solution of cloud PaaS such as Chef or Cloud Foundry Change.In one embodiment, the effect of CAPS is to ensure that movement is carried out effectively, and data is available at new projects' continuation Reason.Attribute may include size, network configuration, SLA, elemental user privilege, backup policy and affinity rule.

Fig. 5 B show can by virtualization and projecting subassembly 221 execute (for example, via to expose management interface for can The order line accessing) to build the exemplary series of orders 510 of virtual data center.Although showing for explanatory purposes Command line interface, but may be in response to user in graphic user interface operating element and execute identical Management Information Base (for example, such as Shown in Fig. 5 C to Fig. 5 D).In this illustration, " establishment data center ' Bob ' " orders the data center for entitled ' Bob ' (being represented by figure 501) creates Virtual Controller.Order " in ' Bob ' upper establishment sub-network ' main ' 192.168.1.0/24 " exists Create network switch Virtual Controller 504 under virtual data center ' Bob ', and another group " establishment " orders in network exchange Machine creates net virtual controller 502, File System Virtual controller 510 and two virtual machine controllers 508 504 times.Gained Virtual data center subsequently via " projection " order and project to data center 520.As described herein, various differences can be adopted Technology virtual data center is projected to typical data center.After virtual data center successfully projects, can perform and " move Dynamic " order so that virtual data center is moved to new typical data center 521.

Fig. 5 C shows exemplary graphical user (the GUI) " chain of command for creating and managing virtual data center Plate (dashboard) ".Liang Ge data center (data center A 553 data center B 554) is shown in GUI window 551 The network topology being coupled together via WAN accelerator.In one embodiment, user can be led in shown network topology Cross and select and drag the pel representing Virtual Controller creating, to edit and to delete Virtual Controller.For example, may be selected in GUI Region 550 in shown Virtual Controller element, and the related configuration number of each of editable and these elements According to.Alternatively, user can directly select Virtual Controller to edit the variable related to controller from GUI window 551.

Also show station state window 552 is had with arbitration (for example, data center's cost), Performance And Reliability with providing The data closed.Under figure arbitration element, the various cost information of user-accessible, include maximum specified cost, objective cost and To each data center using related most current cost.In addition, under figure arbitration element, user may be in response to value at cost Change and specify triggering and action (being referred to as " event " in figure 2d).For example, user could dictate that, if value at cost rises above Specific threshold, then should migrate.

Under graphics performance element, user can check current performance measured value, including network performance, cpu performance and storage Performance.User may also be responsive to change and regulation triggering and action in performance number.For example, user could dictate that, if any specific To less than defined threshold, then data center should migrate the hydraulic performance decline of variable.

Below figure fault management element, the various fault/reliability data of user-accessible, including different types of system System reports to the police (for example, serious warning, main warning, secondary warning etc.).Again, user may be in response to reliability change and Regulation triggering and action.For example, user could dictate that, if the serious quantity reported to the police or mainly report to the police rises above regulation threshold Value, then data center should migrate.

In one embodiment, the management GUI shown in Fig. 5 C provides following functions/feature：

Distributive data center in their corresponding geographical position it can be seen that；

Campaign item illustrates to provide that IaaS service provider of service；

Restriction should projected virtual data center strategy ability, including based on position, cost, time, performance, SLA, layer Level, the strategy replicating；

Alternative data center's (monitored website) should be visible to terminal use；

((for example, IaaS provider, it is just to include performance issue, Cost Problems and availability issue to be clear that existing problems There is provided limited time with regard to its assets, be likely to be of count down timer)) in place of ability；

Plan ahead moves, and may monitor resource in the destination locations available ability of holding；

It is clear that cost in different pieces of information focus instance, and cost is intracardiac in data-oriented (for example, can remain every Hour report mechanism is to carry out facility finance evidence obtaining at the end of the moon, season or year) situation ability；And/or

Ability with regard to the mobile configuration strategy of data center.This may include the alert notice needing to intervene manually, and based on true in advance Fixed maintenance window and dispatch the ability (for example, if arbitration event occurs, in the movement of 2 points of the next morning) of migration.One Kind special it may be the case that, if existing projection is passed due to the fault in data center, move immediately.

Fig. 5 D illustrates in greater detail the hierarchical virtual controller arrangement illustrating in window 551.As mentioned, user Can only be designed in virtual data by selecting and dragging the graphic elements of each of expression Virtual Controller 560 to 599 The heart.In specifically topology shown in Fig. 5 D, gateway 560 is associated with data center A, and gateway 561 is associated with data center B. Network router 563, fire wall 562 and WAN accelerator 564 are positioned directly under gateway 560 according to logic, and another network road It is positioned directly under gateway 561 in layered arrangement by device 582, fire wall 583 and WAN accelerator 581.As illustrated, making two The communicatively coupled special WAN connectors of individual WAN accelerator 564 and 581 can be used for guaranteeing the redundancy between Liang Ge data center And/or standby.In one embodiment, as discussed below, WAN connectors can be used for simplifying transition process (that is, in general When virtual data center moves to new typical data center).

First group of network switch 565 to 567 according to being logically positioned at below the network router 563 of data center A, and Second group of network switch 584 to 586 is according to being logically positioned at below the network router 582 of data center B.Switch 565 One group of Apache Server is couple to local network, this local network is mono- by 568, one group of live load/VM of load balancer First 569 to 571 (for executing process task) and file system 572 form.Switch 566 will be mono- for second group of live load/VM Unit 573 to 576 is coupled to realize cache subsystem (" Memcache "), and switch 567 is by another group of work Load/VM unit 577 to 579 and object memories 580 are coupled to realize data base.

In example shown in Fig. 5 D, one group of assembly of mirror reflection is configured in data center B.For example, switch One group of Apache Server is couple to local network by 584, and this local network is by 587, one group of live load/VM of load balancer Unit 588 to 590 (for executing process task) and file system 591 form.Switch 585 is by second group of live load/VM Unit 592 to 595 is coupled together to realize cache subsystem (" Memcache "), and switch 586 is by another group of work Load/VM unit 596 to 598 and object memories 599 are coupled to realize data base.

In one embodiment, can only pass through to be clicking and dragging on Virtual Controller in shown layer architecture, and add volume Outer data center's element, such as process resource, Internet resources and storage resource.For example, extra work can be loaded/VM control Device processed is added under each corresponding switch to lift disposal ability.Similarly, extra switch can be added on route Under device 563,582, new subsystem is added to data center topology.

In some of embodiment described below, file system 572 and 591 is distributed file system, and it has interior The WAN interconnection (for example, as Hadoop) of crossing between Liang Ge data center put maintains synchronous ability.

As the particular example employing framework shown in Fig. 5 D, network service can be by by load balance device 568,587 And keep the Apache Server group of load balance to provide.These can be in single subnet network.Memcache server can Form the second sub-network to maintain the active cache of their corresponding data bases.Database server group 577 to 579,596 to 598 respectively since shared data memorizer 580,599 operation.

In operation, when URL request by gateway 560,561 enter data center when, URL request by fire wall 562, 583 screenings, and it is then forwarded to load balancer 568,587, described request is redirected by described load balancer One of Apache Server.In this case, the quantity of server can increase under heavy load condition automatically.For example, can limit Determine the scope of server, and trigger can be used for making those server pools extend and shrinks.Being unique in that here, institute Show that framework is actual logical datacenter, it is orthogonal with any given cloud provider service, so that it is inherently Light.

Return to last example, once URL process, movable Apache Server will turn to Memcache server Send out request.If there is dirty bit (for example, data is expired) in Memcache, then in one embodiment, Memcache will respond to stale data (for example, in 200ms) immediately, rather than wait the several seconds to be refreshed.Work as thing When part occurs, Memcache will trigger the data base querying from next group server.During do so, when terminal use is in Qi Liu Look at and click on device when refreshing, they generally obtain latest data.In other words, it is left for terminal use and carry out re-request data general Give the renewal " behind the scenes " of Memcache necessary time.

Fig. 6 shows in an embodiment of virtualization and projection logic 221 for projecting to virtual data center Each layer of cloud provider.Specifically, abstract virtual data center represents that 601 can be using the GUI shown in Fig. 5 A to Fig. 5 D And/or build via third party's user interface 602 (for example, using the GUI designed by third party).In one embodiment, Each object (for example, the Drawing Object 560 to 599 shown in Fig. 5 D) from abstract GUI layer 601 is mapped to virtual bench control A specific controller in device layer 603.Each virtual bench controller includes virtual data structure, this virtual data structure bag Include the data realized needed for hardware foundation part (for example, router, switch, gateway etc.), and be able to access that connecing of data Mouthful.For example, described interface can be realized using representative state transfer (REST) or any other interface model.

One group of virtual bench controller 603 of gained can be via cloud intermediary layer (cloud mediation layer) 604 It is mapped to the corresponding physical unit in institute's data for projection center 605, described cloud intermediary layer can using cloud API (for example, JClouds) realizing.In one embodiment, provide single " plug-in unit " module that virtual bench controller is represented mapping And/or the form that can realize in the resource being provided by cloud provider is provided.Therefore, in figure 6, plug-in unit A is used for inciting somebody to action Virtual data center represent mapping and/or be transformed into cloud provider A, plug-in unit B can be used for by virtual data center represent conversion and Project to cloud provider B.Therefore, when new data center is to its service of intermediary 210 registration, do not need to change underlying virtual Data center represents.On the contrary it is only necessary to existing virtual data center is mapped and/or is transformed into new physics by new plug-in unit Data center.These plug-in units can clothes at (for example, on cloud provider LAN) and/or CAPS 100 in cloud provider guard station Realize on business device.

Fig. 7 shows the other details related to an embodiment of overall intermediary 210.As described above, overall intermediary 210 Including data center 211, described data center comprises the data relevant with each provider, including but do not limit In resource data (for example it is intended that available processing platform, the network platform, storage platform type), performance data (for example, base To measure in the delay related to process task or network communication task), cost (for example, with dollar/sky or other use unit Represent), geodata (for example, indicate geographical position) and reliability data be (for example, based on the important warning in special time period Par).

In one embodiment, multiple application programming interface (API) are exposed, to provide to data center 211 Access, described application programming interface includes cloud provider interface 701, cloud user interface 702, configuration interface 703 and management interface 704.In one embodiment, each interface all includes a set of commands that, (for example, creates note for executing operation on the database Record, deletion record, modification record etc.).Cloud provider can access data center 211 via cloud provider interface 701, Cloud user can access this data center via cloud user interface 702, and data base's configuration is entered via configuration interface 703 OK, managing operation provides via management interface 704.Additionally, it is provided there being messaging bus 705, it allows all cloud users to keep Available resources are carried out with up-to-date checking (for example, provides data center's knot by the queue (being discussed below) listened to cloud user Really).

In one embodiment, management interface 704 is used for carrying out for required the appointing of systems stay execution by CAPS 100 What and all housekeeping operation functions.Can include executing the ability of following behavior by the function that management interface is supported：

Check and change and enliven buyer and the relevant data of seller；

There is provided and limit the accesses control list (ACL) that buyer and seller conduct interviews；

The flux of monitoring message bus 705；

Other computing elements and memory element is made to start or close, to process different loads；

The data center manager client based on SaaS for the management；

Enter the debugging of low-level command row and diagnostic tool；

Closing system simultaneously prepares to move to new data center；

Check all running examples of system, including those run in multiple data centers；

Management standby solution；And/or

For debugging purpose static management client.

In one embodiment, configuration interface 703 allows CAPS 100 to provide the renewal to data center 211 (for example, be data center after examining and/or affiliate data center increases new entry, and remove no longer in service or The data center being no longer necessary to).In non-Partner classification, (for example, non-active consciousness make use of its resource to CAPS 100 IaaS provider) in, when a change, decide whether to provide the renewal to data center 211 by CAPS 100.

In one embodiment, make configuring technical as data center's " region " (used by Amazon) for cloud user not Transparent.For example, can simply region recognition be different data centers.Therefore, can be at a cloud provider (for example, Amazon) A region in using a virtual data center, and one can be used in a different cloud provider (for example, Rackspace) Individual virtual data center.The inevitable outcome of this situation is, cloud user may specify that they wish single provider and different Website/region (for example, using the affinity rule of the affinity showing certain site).

In one embodiment, the function that configuration interface 703 is comprised includes executing the ability of following behavior：

Add/delete/check that IaaS data center (seller) records；

Update static/dynamic seller's record；

Force to push record to buyer；

Create new buyer and optionally create the SaaS infrastructure (using Siaras template) of correlation；And/or

Check and register buyer and the relevant statistical report form of seller.

In one embodiment, cloud provider interface 701 opens to the cloud provider wanting issue available service details. API may be connected to SaaS portal website to be manually entered and to open, from integrated for M2M and automated system.This interface Function may include add/delete/check/update cloud provider record ability.

In one embodiment, cloud user management interface 704 is opened to the cloud user running the virtual data center being managed Put.These systems can run (as SaaS) in cloud itself, also can be used as enterprise application.The purpose of this interface is for quilt The virtual data center of management provides method, to report that the current of virtual data center executes experience.This interface can be embodied as A kind of closed system, described closed system is only passed through just can be activated by the management software that CAPS 100 is supplied to cloud user (that is, client can not directly access this interface).In one embodiment, the function included in cloud user interface include execute with The ability of lower behavior：

The update status of virtual data center performance observed by report；

Any outage observed by report special services provider；

Report that (reported by for example, incompatible between intermediary system or shortage to be configured in virtual data center for any fault The function of accusing or available resource)；And/or

CSAT method for reporting and Trouble Report mechanism are provided.

In one embodiment, overall intermediary is responsible for cloud user and carries out DNS entry redirecting.So, can be right immediately The migration of data center is updated.In addition, an embodiment of overall intermediary 210 is designed to support extension.Specifically, Any interface needing multiple client all supports outward expansion framework.

As described above, virtualization and projecting subassembly 221 can realize typical data center by being converted into virtualization expression Form needed for (i.e. plug-in unit, abstract data is converted to the spendable form of physical resource of data center by it), or by straight It is connected on execution virtual data center in cloud provider, virtual data center is projected on typical data center.

Fig. 8 shows latter event, and it utilizes the embodiment of Full-virtualization, and embodiment is included in virtual data The heart covers 800, and described virtual data center covers and runs on multiple virtual machine 821-826 that general cloud provider 830 provides. Therefore, this embodiment includes virtualized cloud layer again completely, wherein can be by each assembly 801- of virtual data center 800 806 project to any cloud provider (from cloud provider the most general to cloud provider the most complicated).Specific shown in Fig. 8 In example, each assembly runs on the different VM that general cloud provider 830 exposes.Specifically, virtual gateway 801 is in VM Run on 821.Three different kernel VM 802,804,805 (comprising virtual kernel) transport respectively on VM 822,824,825 OK.As illustrated, operating system or other software image 811,812 and 813 can be upper in kernel VM 802,804,805 respectively Face executes.Virtual switch 803 runs on VM 823, and Virtual File System 806 runs on VM 826.

As shown in figure 8, forming each of virtual component 801-806 of virtual data center 800 all can use the 2nd layer Between tunnel protocol (L2TP) tunnel, SSL (SSL) tunnel or another kind of security procedure, agreement to be communicating, thus in assembly Between create secure tunnel.In addition, as illustrated, virtual gateway 801 communicably by other assemblies 802-806 via VM The public IP interface of 821 offers is couple to public network.

It is coated with multiple benefits using virtual data center.Firstly, since not needing to carry out any conversion, therefore can be by void Intend data center's covering to be seamlessly deployed on any typical data center that consistent VM can be provided.Can be in data center services Each grade of layer on apply SLA evenly and be securely distributed.In addition, it is possible to provide the bigger of real data center controls and visible Property, thus over time provides in a more uniform manner.For example, each entity (for example, vGateway, KVM, vSwitch, VFileSystem etc.) may include agency, with continuous measurement performance and cost.

As described above, the file system 510,572,591 realized in one embodiment of the invention is distributed document System, it has and maintains synchronous built-in capabilities by WAN interconnection between two (or more) data centers.Fig. 9 illustrates One such embodiment, described embodiment is included in the first virtual data utilize in the cloud provider 900 file system 920 The heart 910 and the second virtual data center 911 that related file system 930 is utilized on cloud provider 901.File system 920 He Each of 930 all include distributed file system engine 923 and 933, and described distributed file system engine is respectively used to Make the local part 921 of file system 920 synchronous with the remote portion 932 of file system 930, and be used for making file system 930 local part 931 is synchronous with the remote portion 922 of file system 920.Basis after realizing synchronization, to file system 920 Ground part 921 is made any change and is all automatically reflected in the remote portion 932 of file system 930, and to file system The local part 931 of system 930 is made any change and is all automatically reflected in the remote portion 922 of file system 920.One In individual embodiment, " local " assembly 921,931 of file system be by corresponding virtual data center 910,911 locally created, Editor and/or those assemblies otherwise accessing.By contrast, " long-range " assembly be by different virtual data centers 910, 911 those assemblies creating, editing and/or otherwise access.

In one embodiment, distributed file system engine 923,933 is Hadoop distributed file system (HDFS) Engine, and local part and remote portion are embodied as Hadoop node.HDFS is a kind of portable literary composition of distributed expandable Part system, it stores big file on multiple machines and realizes reliability by across multiple main frames replicate datas.Therefore, Hadoop example does not need RSID to store on main frame.Back end can be talked mutually, so that data rebalancing, mobile copy And maintain the data of higher duplication.In embodiment shown in Fig. 9, Hadoop agreement can be used for making file system 920,930 Respective local part 921 and 931 is synchronous with remote portion 932 and 922.It should be noted, however, that the ultimate principle of the present invention is not It is limited to any specific distributed file system agreement.

In one embodiment of the invention, distributed file system (as described above) is used for simplifying data center Transition process.For example, as shown in Figure 10, if user selects to move to Xin Yun provider 902 from cloud provider 900, then one Denier completes new projection (for example, using virtual data center technology as herein described) so that it may use distributed file system 943 The base file system 940 to fill virtual data center 912 for the agreement.For example, in Fig. 10, can be from cloud provider 901 The remote portion 932 of existing file system 930 fills the local part 941 of file system 940.Similarly, can be from cloud provider The local part 931 of 901 existing file system 930 fills the remote portion 942 of file system 940.In one embodiment, It is not necessary to be filled up completely with the local part 941 of file system 940 and remote portion 942 before cloud provider 902 is reached the standard grade Full content.On the contrary, can create run duration (after virtual data center 912 is reached the standard grade) and fill locally distributed literary composition Part system node 941 and remote distributed files system node 942.When receiving at local node 941 or remote node 942 During request of data (local non-availability), this data can be fetched from remote node 932 or 931 respectively, thus in run duration filling Local node 941 and remote node 942.Therefore, virtual data center 912 ratio in cloud provider 902 can be made previously to implement (wherein, need all data of file system 940 in advance) in mode more effectively to start.

Figure 11 A shows another embodiment, and described embodiment includes cloud provider 1100 and is currently running in virtual data The heart 1110, described virtual data center is couple to the file system 1120 of distributed file system engine 1130 management (as previously As in embodiment).In this embodiment, shadow storage system 1101 is used for storing virtual data center file system 1120 Shadow copy 1121.It is communicatively coupled to the distributed literary composition of the distributed file system engine 1130 of virtual data center Part system engine 1131 is configured to maintain the synchronous shadow copy 1121 of file system 1120.In one embodiment, when right When local file system 1120 makes change, distributed file system engine 1130-1131 coordinates with shadow file system Those changes are reflected in system 1121.The difference of this embodiment is that shadow memory 1101 itself is not can be by terminal use The data center accessing.On the contrary, this shadow memory is only used for shadow storage.

In one embodiment, shadow storage system 1101 is used for simplifying data center's transition process.For example, as Figure 11 B Shown, if user selects to move to Xin Yun provider 1102 from cloud provider 1100, then once completing new projection (example As using virtual data center technology as herein described) so that it may be filled using the agreement of distributed file system engine 1132 The base file system 1122 of virtual data center 1111.For example, in Figure 11 B, can be from being stored in shadow storage system 1101 On file system 1121 shadow copy filling file system 1122.

As above-described embodiment, it is not necessary to completely by file system 1121 before cloud provider 1102 is reached the standard grade Full content is filled into file system 1122.On the contrary, can create in run duration (after virtual data center 1111 is reached the standard grade) And fill distributed file system node 1122.When receiving request of data (local non-availability) at file system 1122, This data can be fetched from shadow file system 1121, thus filling file system 1122 in run duration.Therefore, virtual number can be made More effectively start than in previously embodiment in cloud provider 1102 according to center 1111.

As illustrated in fig. 12, in one embodiment, cloud analysis and projection service 1000 can be respectively in different data centers At 1200 and 1210, the gateway apparatus 1250 and 1251 using its own are realized.Gateway 1250 is subsequently used in execution number Set up secure connection according to when migrating in center between data center, such as VPN (virtual private network) (VPN) connects.In an embodiment In, VPN connects the WAN accelerator link including buying, and such as provides duplicate removal compression algorithm (for example, Riverbed) those.

Two tenants are shown in data center 1200：There is virtual data center 1230 and the rent of file system 1220 Family 1201, and there is the tenant 1202 of virtual data center 1231 and file system 1221.There is virtual data center 1232 It is located in data center 1210 with another tenant 1203 of file system 1222.As used herein, " tenant " is to have registered herein The cloud user of described virtual data center service.In the example shown, special VPN is connected to the virtual number of tenant 1201 Move to data center 1210 according to center 1230 from data center 1200.In this case, because VPN connects (by CAPS 100 purchases) it is private link between data center, bring the extra of relevant data center migration thus without to tenant 1201 Cost.

As shown in Figure 12 B, in one embodiment, except providing local gateway at each of cloud provider A-C place Outside device 1300-1302, cloud analysis and projection service 1000 can build/buy network structure (for example, dedicated network basis sets Apply/backbone network etc.), described network structure includes other gateway/router device 1310-1313, for supporting each to carry For the high-speed secure interconnection between business.

In one embodiment, cloud analysis and the projection service 1000 all provider's connection tables as indicated in fig. 12 c of maintenance 1290, to determine that whether there is special high speed between the data center of any two cloud provider connects.Shown in Figure 12 C Specific example in, provider 1-3 is all carried out mutually via dedicated network infrastructure (for example, maintained/bought by CAPS 100) Connect, and cloud provider 4 and 5 does not have any such connection.In one embodiment, cloud analysis and projection service 1000 and/ Or information sheet 1290 when selecting engine 220-222 can recommend or select providing data center.For example, if data center 3 and 5 There is similar cost, Performance And Reliability feature, but data center 3 is couple in current data via special high speed connection The heart, then select engine can have precedence over data center 5 and recommend to move to data center 3.

As described above, overall intermediary 210 can dynamically be updated using the feedback from cloud provider, described feedback may include Cost, performance and/or reliability update.Figure 13 A shows an embodiment, and wherein performance and/or reliability update are by spy The 1320-1321 that acts on behalf of determining execution in the virtual data center 1301 of tenant dynamically provides.In this particular embodiment, will act on behalf of 1320-1321 inserts in each live load 1310-1311 of execution just in the resource of virtual data center 1301 respectively.Generation Reason 1320-1321 monitors implementation status and the collect performance data of corresponding live load 1310-1311.For example, act on behalf of 1320-1321 can measure the time that live load 1310-1311 completes needed for configuration processor code, and/or can check void Intend the other assemblies (for example, file system 1325) in data center and measure the time receiving from each assembly used by response. Subsequently this information can be reported back to overall intermediary 210 via gateway 1330, and for calculating the normalizing of cloud provider 1300 Change performance measurement.

Figure 13 B shows another embodiment, and wherein individually data collection effort loads 1360 and virtual data center Other live loads 1350-1352 executed in parallel in 1301, with collect performance data and/or reliability data.Data collection Live load 1360 can check other live loads 1350-1352 and other data center component (such as, file system 1353) And measure the time quantum (time longer expression performance is relatively low) receiving used by response.Data collection effort load 1360 also may be used Calculate the time quantum executing needed for the program code of itself (in program code, for example, to insert label or execute other programs Tracking technique).Hold as another live load because described data collection effort is supported in the resource of cloud provider 1300 OK, the performance therefore executing correlation with itself represents the performance of other live loads 1350-1352.Data collection effort loads Subsequently results of property is fed back to overall intermediary 210 via gateway 1330.

In another embodiment, gateway 1330 itself can from live load 1350-1352 collect performance data (for example, Check live load as mentioned above), and the data obtained is fed back to overall intermediary 210.

In shown in Figure 13 A to Figure 13 B two embodiment, act on behalf of 1320-1321 or data collection effort load 1360 The reliability of virtual data center also can be monitored.For example, if being not received by for the response checking after the stipulated time, Can determine that just measured assembly is unavailable.Subsequently via gateway 1330, this reliability information can be transferred to overall intermediary 210.

As described above, in one embodiment, CAPS 100 framework can be used as online marketplace, can make for purchase and sale The data center services being built with CAPS 100 framework.Additionally, market is not limited to be bought by cloud user and by actual cloud provider Sell.But, in one embodiment, any user or entity all purchase and sale data center can take in opening the markets Business.Therefore, a particular cloud provider can buy data from another cloud provider (including virtual provider as discussed below) Center service, thus meet demand.Conversely, cloud user can be to another cloud user or in another cloud provider sales data Central server.For example, a particular cloud provider can sell this data center clothes in the earlier month of estimated use data center services Business (for example, June sells the data center that will use in December/January).By using provided by CAPS 100 Open the markets, commercially available these of another user or cloud provider service and subsequently sell these services again and (for example, get a profit or lose Damage, be specifically dependent upon the market price when December/January for this service).So, CAPS 100 just establishes in data The forward market of central server.

Figure 14 illustrates how to be configured to realize by overall intermediary 210 data central database 211 such online Market.As it was previously stated, data center 211 comprises the respective state-of-the-art record of cloud provider to CAPS 100 registration, institute State record to include resource data, performance data, cost data, geographic position data, reliability data and may have with cloud user Any other data closed.Record 1401 is related to all current informations of Amazon service (AWS) and includes described All current informations.By contrast, record 1402 represents that having been for certain stipulated time section have purchased AWS data with certain price User's (or another cloud provider) (being identified as virtual AWS (vAWS) data-base recording) of center service.Return to above-mentioned example, This user several months ahead of time may have purchased the data center services of December/January from AWS, but is not intending to using this A little data center services (or perhaps for only being bought using a part of intention, or anticipate after buying a period of time Know the service that he/her is not needed all to be bought).In one embodiment, bought data center services future (or Currently) user of authority can register the availability of these services in the data center 211 of overall intermediary 210.Then Overall intermediary 210 may be in response to include these services from the data base querying of various selection engine 220-222 generation.If this The cost of a little services is less than Vehicles Collected from Market valency (every other factor is identical), then select engine 220-222 will recommend/select this A little services, these services are better than those services being provided with Vehicles Collected from Market valency (it is assumed that seller is buied with lower price, then this can Seller is made to make a profit).

Figure 14 shows and the related other details that communicate selecting between engine and overall intermediary 210.Specifically, may be used Query argument 1405 is sent to database query engine 1400, described data base from each of selection engine 220-222 Query engine is subsequently used parameter to inquire about data center 211.For example, engine 220 is selected may to be pointed to knob About interested with Japanese data center；The data center that selection engine 221 may be pointed to California is interested；And select Select engine 222 may be pointed to Europe data center interested.As response, database query engine can perform an inquiry (or a series of inquiry), and obtain all data centers meeting required standard from data base 211.In one embodiment, tie Fruit is that above-mentioned " candidate " data center (that is, meets the data center of certain minimum primary standard；In this case, it is based on ground The data center of reason position).

In one embodiment, the result of database query engine 1400 is provided as the entry of queue 1410, and selects Select each of engine the entry of queue 1410 is read out/filters.In one embodiment, implement Producer/disappear The person's of expense framework, wherein database query engine 1400 are used as Producer (new entry is write into queue), select engine 220-222 to use Make the consumer " listener " of queue (otherwise referred to as) of queue.Return to above-mentioned example, select engine 220 will only from in knob Those entries are retrieved in about related with Japanese data center queue 1410；Select engine 221 will only from California The related queue 1410 of data center in retrieve those entries；And select engine 222 by only from European data center's phase Those entries are retrieved in the queue 1410 closed.Each selects engine 220-222 subsequently to can perform filtration/weight behaviour as above Make, to filter candidate data center further, thus obtaining the recommendation favourable to cloud user or selecting (for example, based on cost, property Energy, reliability etc. are filtered).Although being not explicitly depicted in fig. 14, each selects engine can generate inquiry in queue, Search for relevant those entries (for example, for selecting engine 220, be New York and Japan) only to retrieve with it.

Embodiments of the invention may include various steps set forth above.These steps can be presented as make general processor or Application specific processor executes the machine-executable instruction of some steps.Or, these steps can be by comprising for executing these steps Firmware hardwired logic specific hardware components execution, or held by any combinations of the computer module programming and custom hardware components OK.

The element of the present invention also may be provided as the machine readable media for storing machine-executable program code.Machine Computer-readable recording medium may include but be not limited to floppy disk, CD, CD-ROM and magneto-optic disk, ROM, RAM, EPROM, EEPROM, magnetic card or light Card, or be suitable for storing the other kinds of medium/machine readable media of electronic program code.

In entirely described above, for illustrative purposes, set forth many specific detail thoroughly to understand the present invention. It will be apparent, however, to one skilled in the art that can be in the situation of some in not having these specific detail The lower enforcement present invention.For example, it is obvious to those skilled in the art that functional module as herein described and side Method can be implemented as software, hardware or any combination of them.In addition although describing in the case of mobile computing environment herein Some embodiments of the present invention, but the ultimate principle of the present invention is not limited to mobile computing embodiment.In fact, in some enforcements In example, any kind of client or peer-data processing meanss can be used, including (such as) desk computer or work station meter Calculation machine.Therefore, scope and spirit of the present invention should be judged according to appended below book.

Claims (6)

1. a kind of equipment, including：

For managing the intermediary component of data center information, described intermediary component receives the renewal comprising described data center information And inquiry is responded to provide described data center information；

Select engine, described selection engine is cloud user execution number based on the data center information reading from described intermediary component Select according to center and/or recommend, wherein said data center selects and/or recommends the data center based on terminal use's defined Require and/or preference and make；And

Projecting subassembly, once select data center, described throwing from the described described selection selecting done by engine and/or recommending Shadow assembly and selected data center interact and transmission data is to represent data center described in user configuring.

2. a kind of method for selecting data center, including：

Generate the inquiry to the information relevant with data center, the data storage comprising data center information is pointed in described inquiry；

Using described inquiry, in described data storage, execution is searched and is generated one group of candidate data center；

Based on the requirement by described cloud user's defined come excellent to one or more of described candidate data center for cloud user First sort, described require to include indicate data center use the cost requirement of cost threshold value, include performance threshold performance Require and/or include the reliability requirement of reliability thresholds；And

It is that described cloud user selects one of described data center based on described priority ordering.

3. a kind of virtual data center equipment, including：

Virtual data center layer, described virtual data center layer includes multiple virtual bench controllers and defines described virtual bench The data of relation between controller；

Wherein each described virtual bench controller represents physics data center resource and its relevant configuration, and combine is multiple Virtual bench controller forms virtual data center；

Cloud intermediary layer, described cloud intermediary layer in response to described virtual data layers are projected to the order of typical data center and by institute State multiple virtual bench controllers and be mapped to the related data center resources on typical data center.

4. a kind of equipment, including：

Graphic user interface (GUI), described graphic user interface includes the multiple objects that can be selected by terminal use and edit；

Virtual bench controller layer, described virtual bench controller layer includes multiple virtual bench controllers, each virtual bench Controller represents physics data center resource and its relevant configuration, and the plurality of virtual bench controller combining forms void Intend data center；

The each of the multiple described object of wherein said GUI represents and to map to virtual bench controller in the layer specific virtual Setup Controller；And

Wherein user can edit data by selecting the described object related to the described virtual bench controller of those resources Center resources and the relevant configuration of described data center resource.

5. a kind of equipment for migrating data center, including：

Data center's projecting subassembly, described data center projecting subassembly causes the migration and management data center between；

First example being configured in the first data center of distributed file system engine and distributed file system engine The second example being configured in the second data center, described first example of described distributed file system engine and described Second example represents user and keeps same being stored between the file of described first data center and described second data center Step；

Described data center projecting subassembly by open described distributed file system engine in the 3rd data center 3rd example leads to the migration from described second data center to the 3rd data center, and described the of described distributed file system Three examples fill file from described first data center and/or described second data center to described 3rd data center, described Distributed file system allows described data center projecting subassembly being stored in described first data center and described second data The all files at center make described second data center offline before all copying described 3rd data center to.

6. a kind of equipment for realizing data center services market, including：

The intermediary component of management data center information, described intermediary component receives and comprises the renewal of described data center information right Inquiry responds to provide described data center information；

Data base, described data base can be accessed by described intermediary component, data center information described in described database purchase, wherein Each record in described data base includes typical data center provider or the data of virtual data center provider,

Each record of typical data center provider, each record of described typical data center provider is included described in identification The money of the resource that the first cognizance code of typical data center provider, regulation are provided by described typical data center provider Source data and the cost data specifying the cost related to using described resource；And

Each record of virtual data center provider, each record of described virtual data center provider includes identification and purchases Buy within the regulation persistent period to specify cost to use second knowledge of the user of the right of the resource of typical data center provider Other code, described record identifies described typical data center provider further and is carried by described typical data center provider For resource；And

Select engine, described selection engine is based on the described thing of inclusion reading from the data base that can be accessed by described intermediary component Reason data center record and described virtual data center record data center be recorded as cloud user execution data center select and/ Or recommend, wherein said data center selects and/or recommends the data center based on terminal use's defined to require and/or preference Make.