Cellopoint Cloud Series 1: the Past, Present, and the Future of Cloud Computing

2010/4/20

What on earth is the currently hottest Cloud Computing? What is its difference from the Grid Computing? This article will take you to the origins, conceptions, and related applications of Cloud Computing. You might have heard another noun, Grid Computing, before Cloud Computing was stirred up. Many people consider Grid Computing & Cloud Computing very much alike. In fact, there is no strict segmentation between the two concepts. They are both considered the concepts derived from Distributed Computing.

Grid Computing VS Cloud Computing
Grid Computing:
It is made of the virtual computing cluster by using the un-used resources (CPU resources & Disk Storage) from a large number of heterogeneous computers (usually called Desktops), and it provides a structure for solving massive computing problems. Grid Computing focuses on the abilities of cross-domain computing support. With Parallel Computing applied, it focuses on the full-use of resources between and across the companies to jointly solve the tough computing tasks。
Cloud Computing:
It is a kind of dynamically scalable computing. The basic concept is to divide the task of computing into several processes. After they are processed and analyzed by the servo group (cloud hosts) distributed over the Internet, the outcomes will be returned to the end-users. Although Cloud Computing originates from Parallel Computing, it is not away from the concepts of Grid Computing. But, Cloud Computing focuses more on the processes of data.

Mainstream Cloud Technologies:

MapReduce :
It is the key technology that Google applies to Cloud Computing, which allows developers to develop more programs that process massive data. First, it divides the data into unrelated segments through the Map program for a large number of computers to process. Results are further gathered and integrated through the Reduce program. Then it outputs the outcomes required by developers.
Hadoop:
Hadoop is an open-source program inspired by the Google Cloud Structure. The structure of Hadoop is implemented with the concepts proposed by the Google BigTable and the Google File System. It is written in Java, which can provide a Distributed Computing environment for massive data. But the Distributed File System used is different from Google’s. Yahoo is the main contributor and user of the program.。

Service Patterns of Cloud Computing
The application of Cloud Computing usually provides the clients through the Internet with information technologies, including computation, storage, and bandwidth, in a virtual form of “services”. Through Cloud Computing, users only have to take the services as Black Boxes and input the actions required. They don’t have to know the operations inside the Boxes. They only have to wait for the outcomes returned.
Three patterns based on service categories:

1. Software as a Service, SaaS
The SaaS is a pattern of acquiring the software deployment through the Internet. It provides the company with the Software on Demand from the front-end office applications, such as Email and word processing to the back-end data analysis, customer relationship management, business process management, and human resource management. Representatives are Google, Salesforce, Microsoft, etc.

2. Platform as a service, PaaS
The PaaS is a kind of combination of Servo Hosting Platform and a Virtual Solution. Users don’t have to construct the hardware hosts and the operating systems by themselves. Through the rented Internet, PaaS service providers provide the Virtual Hosting Platform, which saves software & hardware maintenance and labor & time management. Through the PaaS, software providers can focus on the software development and accelerate function deployment online. Well-known developers are Amazon web service, Google App Engine, etc.

3. Infrastructure as a service, IaaS
IaaS makes the IT infrastructure kind of service. The company outsources the structure required within the company to the IaaS contractors. Compared with the costs of ordering hardware, software, storage, power, and the bandwidth for construction of traditional computer room, the company can acquire the IT resources more efficiently by paying per use. The concepts of the Private Cloud & the Hybrid Cloud are extensions of the IaaS。Private Cloud makes the exterior resources interior within the company through the VPN; Hybrid Cloud, which integrates cloud services from different providers more flexibly, combines the Public Cloud/SaaS and the Private Cloud. Sensitive data are served by the Private Cloud while non-confidential data are served by the Public Cloud of lower costs.

More and more suppliers are investing in the cloud services, and that means the Cloud service Market has become the trend for the future. The rise of the market means that the company can lower the construction costs of information services and that it can focus on the core of its operations to improve efficiency and competitiveness. However, Cloud services also bring about many problems, such as security apprehensions, whether or not the Service Level is sufficient for dealing with the daily operational requests from the company, the compatibility with the existing systems, etc. In the presence of the Cloud security problems, the next article will take you to the new technologies and its applications developed by information security providers.