Cloud Deployment And Service Models

Published Date: 02 Nov 2017

Abstract: - Cloud computing is an on-demand, pay-as-you-use model that creates a flexible and cost-effective means to access compute resources (services). Cloud computing is recognized as a revolution in the computing era, the key idea of which is to form a computing pool that can distribute its resources based on the needs of the users. A cloud provider utilizes the virtualization technologies in their systems and customers are charged based upon the amount of resources (such as computers, infrastructures, data storage and application services) used or reserved.

Performance monitoring is an integral part of the cloud maintenance. Approximating the performance of the physical infrastructure is somewhat possible based upon the utilization of the resources but the virtual environment poses a difficulty due to shared and dynamic nature of resource allocation.

In the present scenario, the big cloud giants are continuously working upon the performance of the cloud-based applications to make them more efficient

for the customer to work with. But still, there exist a number of performance issues which if worked upon can enhance the performance of the cloud to a great extent. So, the dissertation proposal would like to define the tools and techniques for the fast retrieval of information in cloud computing with the objective of developing a new methodology in cloud computing for data fetching of the components in a robust manner using different tools like MATLAB and C/C++.

1.Introduction

1.1 Cloud Computing

Cloud computing is an on-demand, pay-as-you-use model that creates a flexible and cost-effective means to access compute resources independent of the location. Cloud computing is still in an early stage and the standards are very important. Cloud computing is rapidly gaining height, as evidenced by implementation and growth of commercial cloud platforms. Cloud computing is recognized as a revolution in the computing era, the key idea of which is to form a computing pool that can distribute its resources based on the user’s needs.

Figure 1.1 Broader View of Cloud

Thus, in the broadest terms, a ‘cloud’ is an elastic execution environment of resources involving multiple stakeholders and providing a metered service at multiple granulites for a specified level of quality (of service). [10]

Today, the leading companies, such as Amazon, Google, IBM, Microsoft and Salesforce.com offer their cloud infrastructure for services.

Figure1.2 Cloud Deployment & Service Models

In brief,

SaaS: cloud-based applications

IaaS: processing and storing data

PaaS: developing, testing and running applications for cloud

1.2 Why companies don’t prefer to maintain their own server systems?

Inflexible computing power

Incalculable maintenance costs

Insufficient data security

Performance, capacity and scalability issues

Limited data accessibility

1.3 Risks

Lack of transparency

Reliability and performance issues

Vendor lock-in and lack of application portability or interoperability

Security and compliance issues

Cloud-service provider viability

High value cyber-attack targets

Risk of data leakage

1.4 Improving cloud

Reliability might be improved if we make use of multiple redundant sites

Security could be enhanced by centralization

1.5 Performance

It is defined as the quality of service (QoS) provided by the service provider to the client or the customer over the network. The performance of a network is calculated in different ways because each network is different in nature, design and the topology used. The performance in case of the circuit-switched networks might be calculated using the queuing models which help the network planner to analyze the performance of the network and make necessary changes and adjustments over the network and the performance metrics that is used so as to improve the performance and thus provide the customer with the best value-for-money service.

3. Present Work

3.1 Scope

The performance analysis based on two leading cloud computing platforms with the traditional web-based server is obtained using the httperf measurement tool and Planetlab testbed which is analyzed. The two performance metrics studied for Google App were Round Trip time (RTT) and Network Throughput with data size and number of requests on the planetlab node. It is analyzed that the throughput of Google App Engine strongly depends on the data size. [5]

This explains the fact that the throughput is inversely proportional to the data packet size. Cloud computing is one of the most innovative techniques today that redefines the way of communication. The above performance analysis of the two big cloud giants is introduced. The above analysis help us conclude that it is possible to get reasonable performance in Google App Engine if we wisely choose the data size while in case of Amazon, we need to look into the infrastructural architecture so as to enhance its performance. The above analysis took into consideration a limited number of parameters to compare the performance of clouds with the traditional web server. [1][4]

Thus, if we somehow try to reduce the data size that is sent over the network or a cloud-based application, the throughtput can be increased thereby further improving the performance of the cloud-based applications. So, the proposed study will be much more elaborative, taking into account few more parameters like data loss rate, jitter and page load delay.

Then, the tools may be used to make a comparative study about the performance of the application. Thus, an algorithm can be devised that helps to transmit small data units or packets at a much higher speed over a network thereby minimizing jitter and the page load delay. Hence, the development of a new methodology in cloud computing for data fetching of the components in a robust manner needs to be done.

3.2 Problem Formulation

In the present scenario, the big cloud giants are continuously working upon the performance of the cloud-based applications to make them more efficient for the customer to work with. But still, there exist a number of performance issues which if worked upon can enhance the performance of the cloud to a great extent. So, the basic concern is to overcome the above said problem (That is to enhance the performance of retrieval of the information from cloud) by either improving the architecture of the cloud-based applications or devising an algorithm that works upon the parameters of the performance metrics to improve the overall all performance of the cloud-based application and provide the customer with the QoS he is paying for.

3.3. Objectives of the Study

Review of performance aspects in cloud computing

Comparison of various performance models/ cloud providers on the basis of attributes

Designing an algorithm to improve the performance of cloud-based applications in terms of fast retrieval of components.

Solution

The proposed problem can be solved by making use of programming language (C/C++) and the Apriori Algorithm (cluster related algorithm). The simple concept of physics is being applied i.e. time is directly proportional to distance if speed of fetching or data retrieval remains the same. Various searching and sorting algorithms will be used that would be retrieved from the database using Apriori algorithm, thus measuring the Euclidian distance and hence calculating the performance in cloud-based application systems.

4. Conclusion

Thus, on an ending note, cloud is an evolving technology that is gaining pace and would change the world communicates. Hence, considering performance to be a key issue in this emerging technology and dealing with it would help define newer ways towards globalization by enhancing the speed of data retrieval and therefore, enhancing the performance of cloud-based application systems.