CN105162826A - Cloud computing multilayer cloud architecture - Google Patents

Abstract

The present invention provides a cloud computing architecture, that is, a cloud computing multi-layer cloud architecture. The invention separates and organically integrates the control layer, data layer, and service layer of the entire cloud, and divides the entire cloud layer into large clouds and medium clouds. Organize the architecture with Xiaoyun. The big cloud is deployed in the cloud architecture center, which has a unified management view of the entire architecture, coordinates monitoring and management of the entire cloud layer, and provides resource push services for Zhongyun and Xiaoyun; Zhongyun is deployed in each physical area. Center, Zhongyun provides distributed storage for the data resources of the entire cloud layer and ensures the consistency and fault tolerance of data resources; Xiaoyun provides high-quality services for end users, and is flexibly deployed in the center of each user area to provide high bandwidth and low latency , Highly concurrent services, the management of each cloud layer in the entire architecture is compact, and the mutual cooperation can effectively improve the service computing capability of the cloud computing architecture and increase the scalability of the cloud layer.

Description

A cloud computing multi-layer cloud architecture

technical field

The present invention relates to the field of cloud computing cloud layer management, and more specifically, relates to a cloud computing multi-layer cloud architecture.

Background technique

Cloud computing is a computing mode of resource sharing. This mode provides available, convenient, and on-demand network access, and supports payment according to usage. The resources provided by cloud computing can be quickly provided, and the advantage is that only a small investment is required. administrative work and minimal interaction with service providers.

At present, the cloud computing architecture usually adopts a simple binary model of cloud centralized management and user pull-down. This architectural model restricts the ability of the cloud to provide large-scale concurrency, high-quality services and active push. In applications with high resource usage, there is a huge performance bottleneck.

With the rapid development of cloud computing, the demand for cloud services is increasing, various businesses are becoming more and more complex, and users have higher and higher requirements for service quality. The service capability and scalability of the existing cloud computing platform can no longer meet the growing needs. In the traditional architecture, the space for performance optimization of cloud services has become smaller and smaller. Cloud service clusters with traditional architecture occupy a large area, consume high energy consumption, have limited service capabilities, and poor user experience have become prominent problems, and currently there is no thorough solution.

Contents of the invention

The present invention provides a cloud computing multi-layer cloud architecture to solve the problems of insufficient service capability and low scalability of the cloud computing architecture.

In order to solve the problems of the technologies described above, the technical solution of the present invention is as follows:

A multi-layer cloud architecture for cloud computing, including large cloud, medium cloud and small cloud. The large cloud includes several medium clouds, and the medium cloud includes several small clouds. The large cloud is deployed in the center of the multi-layer cloud architecture to monitor and manage the entire cloud layer as a whole It also provides resource push services for Zhongyun and Xiaoyun; Zhongyun is deployed in the center of each physical region, and stores the data resources of the entire cloud layer in a distributed manner to ensure the consistency and fault tolerance of data resources; Xiaoyun is flexibly deployed in each user area The center automatically schedules the data resource requirements of users, and provides users with high concurrency, high data volume and virtual desktop services.

Furthermore, Big Cloud collects the statistical information of Zhongyun and Xiaoyun, provides unified system management and monitoring for Zhongyun and Xiaoyun, and analyzes, coordinates and plans the data resource information of all clouds; Big Cloud manages Zhongyun data The index information of Zhongyun establishes a view of the node status, communication pressure, and fault conditions of Zhongyun. Zhongyun performs distributed storage of data resources and uses closed P2P technology to ensure effective sharing of data in different Zhongyun.

Furthermore, Xiaoyun sends a data resource pull request to Zhongyun according to the data index of Dayun and obtains the required data resources. Xiaoyun provides plug-and-play, unattended self-management, and automatic service services for user terminals, and When Xiaoyun is offline, it can still provide offline services for user terminals in Xiaoyun.

Furthermore, Dayun uses multicast and broadcast push methods to push data resources to Zhongyun and Xiaoyun, and Dayun specifies the range, region and users to receive the push; the data resources pushed include text, pictures, videos, etc. , object data, and the cloud service itself.

Furthermore, the automatic service provided by Xiaoyun is that Xiaoyun provides a unified calling interface for Zhongyun or operators, and different Zhongyun or different operators go to Xiaoyun to operate and access Xiaoyun according to the calling interface, and define Xiaoyun's services.

Further, the automatic service offline service provided by Xiaoyun means that Zhongyun will transfer the entire cloud service to Xiaoyun through the call interface of Xiaoyun, and Xiaoyun will directly provide cloud services.

Further, the number of backups and security levels of the self-defined data resources of Zhongyun, Zhongyun includes several Zhongyun servers, when a server in a new region wants to access Zhongyun, it sends an application for accessing Zhongyun, and after the application is authorized, Servers in this new region share data with other Zhongyun servers in accordance with the Zhongyun agreement.

Further, the large cloud includes several large cloud servers, each medium cloud is connected to at least one large cloud server, and each small cloud is connected to at least one medium cloud server.

Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

The present invention separates and organically integrates the control layer, data layer, and service layer of the entire cloud in design, divides the entire cloud layer into large clouds, medium clouds, and small clouds for architecture organization, and the large cloud is deployed in the multi-layer cloud architecture center , has a unified management view of the entire architecture, coordinates monitoring and management of the entire cloud layer and provides resource push services for Zhongyun and Xiaoyun; Zhongyun is deployed in the center of each physical region, and Zhongyun provides distributed storage for the data resources of the entire cloud layer And ensure the consistency and fault tolerance of data resources; Xiaoyun provides high-quality services for end users, and is flexibly deployed in the center of each user area to provide high-bandwidth, low-latency, and high-concurrency services. The management of each cloud layer in the entire architecture is compact and mutual Working together can effectively improve the computing power of cloud computing architecture services and increase the scalability of the cloud layer.

Description of drawings

Figure 1 is a structural diagram of the three main layers of the multi-layer cloud architecture: large cloud, medium cloud, and small cloud;

Figure 2 is a complete video service process of a real-time video service based on the multi-layer cloud architecture;

FIG. 3 is a schematic diagram of a communication protocol of a multi-layer cloud architecture;

FIG. 4 is a schematic diagram of a pull request initiated by a terminal;

Figure 5 is a schematic diagram of the active push initiated by Dayun;

E in Figures 2, 4, and 5 represents the end of the process.

Detailed ways

The accompanying drawings are for illustrative purposes only and cannot be construed as limiting the patent;

In order to better illustrate this embodiment, some parts in the drawings will be omitted, enlarged or reduced, and do not represent the size of the actual product;

For those skilled in the art, it is understandable that some well-known structures and descriptions thereof may be omitted in the drawings.

The technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

Example 1

As shown in Figure 1, a cloud computing multi-layer cloud architecture, the main part of the architecture is divided into big cloud (101), medium cloud (201), and small cloud (401) according to the control plane, data plane and service plane. The servers (102)(103)(104) in the big cloud are responsible for the unified management of the whole cloud and are the control core of the whole cloud. The Dayun server manages all or more Zhongyun at the same time according to its own geographical location and the geographical location of other Zhongyun servers in accordance with the principle of nearby management. Each Zhongyun needs to be connected to at least one Dayun server, and can also be managed by multiple Dayun servers at the same time.

The cloud architecture can be divided into multiple central clouds (201) (202) (203), and each central cloud covers a certain range of regions, such as the geographical range of a province. Zhongyun and Zhongyun servers are connected to each other, servers in the same Zhongyun are connected to each other, all Zhongyun servers are connected to each other, and all Zhongyun servers can use P2P for data transmission.

The intersection of the data of all Zhongyun servers (301) (302) (303) in the same Zhongyun is a coverage of the data that needs to be used in this area. In order to ensure that when Zhongyun in this area completes the data acquisition, it can completely transfer the copy of the file to Xiaoyun.

A small cloud (401) is subordinate to a medium cloud server (302). Xiaoyun supports self-management features, and can provide direct services to other terminals in Xiaoyun offline. Multiple small clouds (402) (403) may belong to the same Zhongyun server (302) as the small cloud (401), or may belong to other Zhongyun servers, such as (303).

As shown in Figure 2, the implementation process of the multi-layer cloud architecture is shown by taking high-definition video as an example. First, each video file corresponds to a configuration file, which specifies the attributes of the video, such as whether it is real-time video, video type and push conditions. The process starts by defining a configuration file:

S1: Before uploading a video, a configuration file (501) needs to be defined first, and the configuration file identifies whether the video type is a real-time video. After the configuration file is defined, Dayun automatically parses the configuration file, adopts different push methods according to different configuration file parameters, and enters the S2 process.

S2: The administrator uploads the video stream. If it is a real-time video, the video will be transmitted while uploading. If it is not a real-time video, the video will wait for the upload to complete before starting to transmit the video. For real-time video (503), it is necessary to start pushing (506) part of the uploaded video stream immediately during the uploading process, so as to ensure the immediacy of real-time video transmission, so that the terminal can start playing video with the shortest buffer waiting time . For non-real-time video, it is first added to the push queue (504), waits for the push permission, and after getting the push permission, pushes it again (505), and enters the S3 step.

S3: Some Zhongyun's subordinates, Xiaoyun, are providing real-time video services for terminals, while some have not yet provided such services. Different operations need to be performed depending on whether there is a terminal (507) requiring service. If a Zhongyun subordinate has a small cloud that needs to provide services, the small cloud downloads the video from the big cloud, and the download of the video stream starts from the starting offset of the video (511), because the small cloud directly belongs to the big cloud Lowest latency and highest quality of service. Video streams can be buffered faster for short periods of time. The data transmission between the big cloud and the small cloud is multi-threaded transmission. The big cloud can also receive the downloaded data of different blocks from other big clouds. When the big cloud receives different blocks, it will perform multi-threaded data transmission to the small cloud ( 512). Different blocks can be obtained by step S4.

S4: If the small cloud under Zhongyun does not need to provide services temporarily or the cached data is sufficient to provide services for subordinate small clouds, this Zhongyun can negotiate with other Zhongyun to download different blocks of the video in segments (508), These blocks are the video sections that the small cloud terminal that is playing does not need to play immediately, and then according to whether there is a terminal (509) that needs to be served in the adjacent large cloud, transmit data (510) to other adjacent large clouds, and finally realize the small cloud. Cloud Service End User Requirements (513).

As shown in Figure 3, the communication protocol of the cloud computing multi-layer cloud architecture is shown. Dayun (601) provides an operation interface for administrators upwards, and an interface for Zhongyun downwards. The data communication between Zhongyun (606) requires interfaces meeting different requirements for data interaction. Zhongyun provides a calling interface for Dayun upwards, and an access interface for Xiayun downwards. Xiaoyun (615) provides an interface upward to Zhongyun, and a service interface downward to end users. The specific interface features of the multi-layer cloud architecture include the following three parts:

P1: Dayun has a global view of the architecture and abstract information of Zhongyun data and status. The administrator can perform data analysis according to the abstract information in the big cloud, and the analysis needs to call the data analysis interface (603). The management cloud can manage the entire cloud through configuration files, and the management process calls the management interface (602). Dayun provides support for query and state synchronization functions for Zhongyun. The query function is to convert the ID of a certain resource into the specific storage information of the resource through the query interface (604). The state synchronization interface (605) is used for synchronous processing when changes, deletions, and new operations occur in the server nodes of Zhongyun or the entire Zhongyun.

P2: The servers of Zhongyun need to ensure the consistency and backup of data, and the state synchronization between different Zhongyun is carried out through the state synchronization interface (611). When the Zhongyun server completes the work that requires multiple Zhongyuns to complete together, it needs to perform task assignment and step-by-step coordination through the negotiation interface (612). Due to the scalability of Zhongyun, the newly connected Zhongyun establishes a connection with the original Zhongyun through the extension interface (613). data transmission. Zhongyun provides a control interface (607) upward, and Dayun conducts global management of Zhongyun through the control interface. When Dayun actively pushes, the push content is transmitted through the push interface (608). Zhongyun provides a query interface (609) for the lower layer, and Xiaoyun can find the specific resource information corresponding to the resource through the query interface. Xiaoyun can obtain data information from Zhongyun through the pull interface (610).

P3: Xiaoyun provides a unified service receiving interface (616) for the upper layer, so that different middle clouds or operators can directly deploy services to Xiaoyun through this interface. Xiaoyun provides a status query interface (617), through which Zhongyun knows the service status of Xiaoyun and whether it is waiting to receive a new push. When Xiaoyun starts push receiving, Zhongyun pushes to Xiaoyun through the push receiving interface (618). Xiaoyun provides a service interface (619) for the terminal.

As shown in Figure 4, it shows a resource pull request initiated by a terminal. Xiaoyun limits the scope of the required resources (701), sends the filter conditions to the query interface of Zhongyun (702), and Zhongyun judges itself Is it possible to complete this query directly for Xiaoyun (703). If the data information in Zhongyun is incomplete, you need to go to other Zhongyun to synchronize information. At this time, Zhongyun sends a query request to Dayun (704), and after getting a response Return the result to Xiaoyun (705). Xiaoyun selects the required data according to the returned result, and sends the data storage index to Zhongyun's pull interface (706). If the data needed by Xiaoyun has a complete backup in Zhongyun (707), it will be directly transmitted to Xiaoyun (709). If there is no complete backup, it needs to be synchronized with other Zhongyun servers according to the storage index (708), and then Transfer to Xiaoyun.

As shown in Figure 5, it shows an active push process initiated by Dayun. The administrator first performs the push configuration (801), and adds this push task to the task queue (802). Initiated or initiated by a task status update (803). When a task is started (805), if it is a new task, the push interface of Zhongyun is called, and Zhongyun negotiates the push scheme through the negotiation interface. If more than one Zhongyun is required to participate in this push, the Zhongyun coordinator for this push is selected, and the Zhongyun coordinator is responsible for managing this push task at the Zhongyun level (808), and calls Xiaoyun’s push after the coordination is completed The receiving port (810), if the push is successful, and you act as the coordinator, then replace it with another Zhongyun to coordinate (815), and the push task of this Zhongyun is completed. If the push task is interrupted or abnormal, the push status will be uniformly reported to the coordinator, such as the push ratio (812), and the coordinator will uniformly report the push result to Dayun (813), and the unfinished Zhongyun will wait for the wake-up state and delay the push (809), waiting to receive a big cloud command. When the re-push command is received again, the Xiaoyun push receiving interface (810) is called again until all Zhongyun pushes that need to complete this push task are successful, and this push task is completed.

The same or similar reference numerals correspond to the same or similar components;

The positional relationship described in the drawings is only for illustrative purposes and cannot be construed as a limitation to this patent;

Apparently, the above-mentioned embodiments of the present invention are only examples for clearly illustrating the present invention, rather than limiting the implementation of the present invention. For those of ordinary skill in the art, other changes or changes in different forms can be made on the basis of the above description. It is not necessary and impossible to exhaustively list all the implementation manners here. All modifications, equivalent replacements and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the claims of the present invention.

Claims (8)

1. a cloud computing multi layer cloud framework, is characterized in that, comprises great Yun, medium cloud and cloudling, and great Yun comprises some medium clouds, and medium cloud comprises some cloudlings; great Yun is deployed in multi layer cloud framework center, plans as a whole monitor and managment and provide resource supplying service to medium cloud and cloudling to whole cloud layer; Medium cloud is deployed in each physically district center, ensures the consistency of data resource and fault-tolerant to the data resource distributed storage of whole cloud layer; Cloudling is flexibly deployed in disposes center, each user area, the data resource requirements of Automatic dispatching user, for user provides high concurrent, high data volume and virtual desktop serve.

2. cloud computing multi layer cloud framework according to claim 1, it is characterized in that, great Yun collects the statistical information of medium cloud and cloudling, provides unified system management and monitoring, and analyze the data asset information of all clouds, plan as a whole, plan medium cloud and cloudling; great Yun manages the index information of medium cloud data, and set up the view of the node state to medium cloud, communication pressure, failure condition, medium cloud carries out distributed storage to data resource, uses effectively sharing of the data of the different medium cloud of closed P2P technique guarantee.

3. cloud computing multi layer cloud framework according to claim 1, it is characterized in that, cloudling sends data resource according to the data directory of great Yun to medium cloud and pulls and ask and obtain desired data resource, cloudling provides the service of plug and play, unattended operation Self management, automation services for user terminal, and when cloudling is in off-line state, still offline service can be provided for the user terminal in cloudling.

4. cloud computing multi layer cloud framework according to claim 1, is characterized in that, great Yun carries out data resource to medium cloud and cloudling and pushes the propelling movement mode adopting multicast and broadcast type, and the scope of specifying reception to push by great Yun, area and user; The data resource pushed comprises text, picture, video, object data and cloud service itself.

5. cloud computing multi layer cloud framework according to claim 3, it is characterized in that, the automation services that cloudling provides refer to that cloudling provides unified calling interface for medium cloud or operator, different medium clouds or different operators go cloudling to carry out operational access according to this calling interface, the service of definition cloudling.

6. cloud computing multi layer cloud framework according to claim 4, is characterized in that, the automation services offline service that cloudling provides is that whole cloud service is transferred to cloudling by the calling interface of cloudling by medium cloud, directly provides cloud service by cloudling.

7. cloud computing multi layer cloud framework according to claim 1, it is characterized in that, described medium cloud self-defining data resource backup quantity and level of security, medium cloud comprises some middle Cloud Servers, when the server in new area wishes access medium cloud, send the application of access medium cloud, apply for that the server in this new area is according to medium cloud agreement and other medium cloud server to share data after authorizing.

8. cloud computing multi layer cloud framework according to claim 7, is characterized in that, great Yun comprises some large Cloud Servers, and each medium cloud at least connects a large Cloud Server, and each cloudling at least connects a middle Cloud Server.