Gloud-the future of cloud computing

Cloud computing is a very advanced concept and technology. Its scale can be dynamically expanded, its processing capability is super powerful, its storage space is massive, its reliability is high, its resource utilization is high, its versatility is strong, and its cost is extremely low. Because of this, the rapid expansion of cloud computing technology at an unprecedented speed has led to the transformation of traditional IT enterprises. At present, there have been hundreds of new cloud computing innovation enterprises in Silicon Valley, with the momentum of the Internet just emerging in the 1990s S.

Cloud computing is undoubtedly the most successful business computing model so far, but it cannot cure all kinds of problems, and its defects are exactly what grid technology is good:

From the perspective of unified platform, there are no uniform standards for cloud computing. Solutions from different vendors have different styles and are incompatible with each other. In the future, we will certainly develop towards a unified platform; the grid technology was born to achieve dynamic integration and sharing of cross-platform, cross-system, and cross-region heterogeneous resources. In addition, the International grid community has formed a unified standard system and successful application. Grid technology can help achieve interoperability between cloud computing platforms, so as to achieve the integration of cloud computing facilities, so that the future of cloud computing is no longer provided by vendors, but constitute a unified virtual platform. Therefore, it is foreseeable that collaborative sharing between the cloud and the cloud is inseparable from Grid Support.

From a computing perspective, cloud computing manages a low-cost computing resource pool composed of PCs and servers, mainly for loosely coupled data processing applications, for computation tasks that are tightly coupled and cannot be easily divided into multiple independent subtasks, It is very inefficient to use the cloud computing mode because frequent communication exists between nodes; grid technology can be integrated with high-performance computers distributed across different organizations. They are good at handling tightly coupled applications, and many applications are tightly coupled, such as numerical weather forecasts, simulated vehicle collision tests, and Stress Analysis on tall buildings. Such applications are not good at cloud computing. If cloud computing and grid technology can be integrated, they can give full play to their respective characteristics.

From the data point of view, cloud computing mainly manages and analyzes commercial data. The grid technology has integrated extremely large volumes of scientific data, such as species genetic data, astronomical observation data, Earth remote sensing data, meteorological data, marine data, drug data, and population statistical data. If cloud computing and grid technology are integrated, the application scope of cloud computing can be greatly expanded. Currently, Amazon is constantly collecting datasets for public sharing, including human genetic data, chemical data, economic data, and transportation data. This fully demonstrates the needs of cloud computing for these datasets, at the same time, it also shows that the collection method is too primitive.

From the perspective of resource integration, to use cloud computing, various data, systems, and applications must be concentrated in the cloud computing data center, however, many existing information systems need to change the operating mode and migrate to the cloud computing platform, which is not easy. In addition, the data sources of some systems may be far away from the data center, and the data sources are constantly updated (IOT has this feature ), if you want to transmit the data to the cloud computing center at any time, the consumption of network bandwidth is not economical. Therefore, there will be a large number of application systems in the scattered state, instead of concentrated on the cloud computing platform; and the grid technology can achieve integration on existing resources, achieving the effect of "Physical dispersion and logical concentration" Can skillfully solve this problem.

From the perspective of information security, many users worry about hosting their valuable data to a cloud computing center, which is equivalent to losing absolute data control, there is a possibility of being viewed, illegally used, or lost by a third party, so that cloud computing technology is not used. In a grid environment, data can still be stored in the original data center and still controlled by its owner, however, the provision of data access services to the outside world is a "usable, but not all can be taken away" mode, without the loss of data ownership, however, the use of data resources is extended and the utilization is improved. Since data sources are controlled by different owners, they can determine whether each type of data is shared and where it is shared, compared with putting all the data into the cloud computing data center for sharing, it is more conducive to avoiding the spread of sensitive data.

Therefore, cloud computing and grid technologies are complementary, not a replacement. The grid technology mainly solves the problem of sharing various information resources distributed across different organizations, while cloud computing mainly solves the problem of centralized sharing of computing power and storage space. It is foreseeable that cloud computing and grid technology will eventually be integrated. This is the cloud computing tomorrow. The author gave it a name called gloud: gloud = grid + cloud.

The benchmark for the combination of cloud computing and grid technology will be service-oriented architecture SOA (Service Oriented Architecture ). SOA was first proposed by Gartner in 1996. The essence of the SOA architecture model is business modeling-encapsulating all information resources into services to solve inter-business interoperability problems in the form of services. Currently, both grid technology and cloud computing basically comply with Web Services specifications. Web services is one of the implementation mechanisms of SOA. Web Services is a software application identified by Uri (Uniform Resource Identifier. The interface and binding of the application can be defined, described, and discovered in XML-based language. At the same time, the application can directly interact with other software applications through the XML Message Protocol Based on the Internet. Web services are popular mainly because they are simple, standard, cross-platform, and independent from vendors. They can greatly reduce the coupling between architectures and provide service-level integration.

In the SOA framework, both grid services, cloud computing services, and other Web services can easily coexist, share, and interoperate. In such an environment, it is difficult to identify which parts are traditional Web Services, cloud computing, and grid technologies, and users do not need to clearly identify them. Users only need to care about what services they need, such as data processing service, high-performance computing service, enterprise management service, and e-commerce service. As shown in.

Figure cloud computing and grid service integration vision

 

It should be said that the cloud age is quite beautiful!

 

Author: Liu Peng, source: China cloud computing (http://www.chinacloud.cn)