A comparative analysis of cloud computing services
Abstract
Cloud computing has changed the discourse of how human civilization does computing, by as-a-service model. As the underlying technologies evolved, many players came into existence. The as-a-service model has taken over the internet space/market. There are so many vendors who provide software-as-a-service, platform-as-a-service, and infrastructure – as – a-service. Because of there are so many parties offering their cloud solutions, the clients/end-users have plenty of choices available. This study examines some of the major cloud service providers and compares them regarding feature richness, cost-benefit, complexity, learning curve, etc. Though there are so many public and private cloud solutions, (it is not possible to study them all in a single document), yet we are going to focus on only two cloud solutions. The one provider is the public cloud provider, i.e. Amazon Web Services (AWS) and the other one is the Open Source private cloud solution, i.e. OpenStack. Using this study, we will compare each cloud platform, by comparing the individual and similar services. Here individual and similar services mean that the services which are having almost same functionality and serve the similar purpose.
Keywords: Cloud Computing • Public Cloud • AWS • Private Cloud • OpenStack • Comparison
Table of Contents
1.0 Introduction 4
1.1 Background 4
2.0. Purpose of the Study 5
2.1. Aim of the Study 5
3.0. Significance & Contribution to Knowledge 5
4.0. Literature Review 5
5.0. Research Approach 8
6.0. Research Methodology 9
7.0. Timeline 9
8.0. Proposed Thesis Structure 10
9.0. References 10
Introduction
“*-as-a-service” model is the most basic philosophy behind the cloud computing. “*-as-a -service” model is not very new. We, humans, are using it since a long time. For example, traveling in bus, train, airplane, or any other public transport, we need to buy the ticket. The ticket just rents fees for a single seat for a specific duration. So, instead of buying a whole bus, it’s convenient and affordable to just rent a seat in that bus by paying a tiny amount. The similar thing happens, in a case of cloud services. The end-user can rent a variety of computing resources whenever (on-demand) s/he needs it, for any period (depending upon the need). The end-user need not buy the whole data-centre for computing the large set of data which demands hundreds of cores of CPUs and memory. Cloud computing enables the end-user to pay a small rent (most of the time pay-per-hour basis) to use that massive computing power. AWS is a public cloud offering by Amazon, which is the leading platform in current times. Anyone can create an account on AWS and start using it, from a very user-friendly dashboard. Opposite to AWS, which is a public cloud, OpenStack provides end-users/companies to host their private cloud in their premises. The OpenStack is fully open Source and free of cost. Also, it can run on commodity hardware. There are hundreds of companies like, Mirantis, who provides the public cloud offering for OpenStack. Some companies use both solutions to meet their application needs. For example, if some company is having clients all over the globe, almost in every country. There are some regulations in many countries that the financial transactions data which belongs particular to that country, can not cross the territory of their country. In this situation, the company needs to use the hybrid approach in which hosting the payment application on OpenStack and keeping it inside the data center located in that specific country. And serving all the global services from AWS, which is easier to scale up.
1.1 Background
The basic idea of cloud computing is very old. The roots of this idea go into several decades back. In the year of 1957 Professor John McCarthy (a computer scientist), who created time-sharing for IBM mainframe computers, started the quest for the cloud computing [1 (McCarthy)]. In John McCarthy’s era, obviously, it was not known as cloud computing. The word cloud computing came into existence decades later. Cloud computing facilitates the user’s on-demand access to computing resources, in very similar fashion to other basic amenities systems. For example, electricity, natural gas, and water supply. In such cases the clients/end-users/household/businesses pay-as-you-go model. One need not buy resources to use them. One can simply use the resources on rent basis. If one need more electricity then, one can simply use it and pay the extra amount. Hence, companies and individuals need not have to deal with capacity planning upfront and can scale up and down whenever needed [2 (Demystifying Cloud Computing)].
Purpose of the Study
The main intention of this study is to understand or analyze the difference between both types of cloud computing platforms. This study can give an insight into the most basic difference between the two cloud solutions. It touches the topics like cost-effectiveness, complexity, learning curve, specific use-cases, etc. This study can advantage the people who are looking for some cloud solutions, for their use-case specifically. The individuals and companies can both benefit from this study. Even, if you are not planning to use all the services provided by these two cloud computing platforms, still you can use this study to compare only those services, which you think, are vital for your applications.
2.1 Aims of the Study
This study aims to analyze two cloud solutions. Each cloud solution is suitable for specific use. The companies/individuals have completely different set of applications. Some enterprises have data mining application; others may have social websites or some content delivery network. The technical capabilities of different companies are completely different. Some may have a big technical team, while others may have a very small one. The aim of the study tries to give a comparative analysis which may benefit companies/individuals while choosing their cloud platform. It also gives a little overview of how one can use the hybrid approach which includes heterogeneous cloud platforms.
Significance & Contribution to Knowledge
In this paper, we show the comparative analysis of the two cloud platforms, i.e. AWS and OpenStack. It can be used to figure out which cloud platform is best for whom by just comparing the two. This study can advantage to small companies who are in their very early stage and still trying to figure out the difference between these two cloud solutions. It may enable them to make informed decisions. It will also point out which cloud solution would be cheaper in short and long run, for any organization.
Literature Review
Cloud computing delivers a very simple way to have access to the databases, servers, storage and a broad range of computing resources in an affordable manner. The only thing that the end client needs is the internet and a browser. The underlying complexity of cloud computing platform is maintained by some really good engineers so that you need not worry about the hardware. One example of such offering is Amazon Web Services [3 (What is Cloud Computing, AWS)]. The other cloud solution OpenStack completely believes in Open Source. It was first created by the platform provided by NASA. Then after a large Open Source, the community took over, where anyone from anywhere can contribute to it. OpenStack can be used The vision for the long-term for OpenStack is to produce the software needed to create private as well as public cloud solution which can scale up to multiple data centers. OpenStack services give control of large pools of storage, networking and compute resources throughout one or multiple data centers [4 (OpenStack Operations Guide)]. OpenStack can be used in a multi-node environment where each service is designed to be scalable, fault-tolerant and secure. That is too keeping the major application goal in mind. Starting with AWS is easier as compared to OpenStack, because of its abstraction layer which hides out the complexity of underlying hardware computing resources from the end-user. In a case of OpenStack, the end-user needs to manage the OpenStack services, apart from the actual application. OpenStack supports multiple hypervisors, which gives the end-user flexibility to choose one solution over other, which fits his/her/theirs needs. Support for some hypervisors is still in development phase, hence can not be confidently used in production environment. All the hypervisors provide different features and support. This document will also seek the differences among these hypervisors. [5 (OpenStack Supported Hypervisors)]. OpenStack works with open source technologies and some very popular enterprise solutions to create a heterogeneous cloud computing infrastructure. Another approach is if you are planning to use both for some specific use cases. Where these specific use-cases may include the applications which need to be inside your data center and some part of the application may be better scalable over AWS, you can use the hybrid approach. In hybrid approach, one couple the two cloud platforms, so that some part of the application is running on each cloud platform. A hybrid cloud approach combines both the public cloud platform and the on-premise private cloud, such that the resources are most of the time not aware of their location. A common cloud management solution may be used to orchestrate between these cloud computing platforms [6 (Red Hat blog)]. For specific use case of cost comparison, companies provide the cost calculator for your needs. Using this cost calculator, you can get approximate numbers for each kind of solution. One example of such calculator is calculator by Mirantis. This free online TCO calculator enables you to compare the long-term costs of using Mirantis OpenStack versus using Amazon Web Services, in as close to an “apples-to-apples” comparison as possible [7 (Cost Calculator by Mirantis).
The below are some example Architectures, for both kinds of cloud computing platforms. :
Figure 1 – OpenStack example architecture [8 (OpenStack Architecture Design)]
Figure 2 – AWS example architecture [9 (The squirrelbin microservice architecture)]
5.0 Research Approach
Step 1:
List out all the services, that needs to be compared to both cloud solutions.
Step 2:
Create an account on Amazon Web Services (AWS).
Step 3:
Start using the services provided by the AWS, like EC2 (compute), S3 (object storage), EBS (block storage), IAM (identity and access management), RDS (database), VPC (networking), Route53 (DNS), etc.
Step 4:
Create a single/multi-node (depends upon the hardware availability) cluster of OpenStack services.
Step 5:
Use the services provided by the OpenStack, like Nova (compute), Swift (object storage), Cinder (block storage), Keystone (Identity and Access Management), trove (database), neutron (networking), etc.
Step 6:
Create a comparative analysis of all the different services provided by each platform, regarding complexity, cost, learning curve, etc.
Step 7:
Create the initial draft.
Step 8:
Complete the final report.
6.0 Research Methodology
The methodology used for this thesis will be simple. We are going to take each of the services provided by the either cloud computing platforms and then we are going to compare them. For example: In a case of AWS, it’s EC2 which is the counterpart of a Nova in OpenStack. The final product of this methodology will be a matrix table containing the services from each cloud computing platform.
7.0 Timeline
The thesis research will take one year.
Started from 23 February 2017 to 22 nd, November 2017.
Time Management:
Topic selection and getting additional information
7 days.
Collecting and Gathering useful data from previous literature
20 days.
Analyzing the previous Literature
5 days.
Research Design
35 days.
To try out both types of cloud platforms
20 days (10 days each)
To perform experiment and finalize the result
20 days.
To make an initial draft
30 days.
To make Final Report
20 days.
Proposed Structure of Thesis
Table of Contents:
Title Page
Acknowledgement
Abstract
Introduction
Analysis of Travel
Travel Analysis using Geotagged Photos
Literature Review
Introduction
Literature Body
Summary
Methodology
4.1 Data Extraction using Geotagged Photos
4.2 Metadata Processing and Clustering
Objectives
Conclusion
Appendix (If Any)
References
References
McCarthy, “Reminiscences on the History of Time Sharing,” Stanford University, 1983, < http://www-formal.stanford.edu/jmc/history/timesharing/timesharing.html>.
Demystifying Cloud Computing, Qusay F. Hassan, Faculty of Computers and Information, Mansoura University, Egypt. http://www.crosstalkonline.org/storage/issue-archives/2011/201101/201101-Hassan.pdf.
“What is Cloud Computing?”. Amazon Web Services. 2013-03-19. Retrieved 2013-03-20.
OpenStack Operations Guide. https://docs.openstack.org/ops-guide/. updated: 2017-04-01 14:46
OpenStack Supported Hypervisors. https://wiki.openstack.org/wiki/HypervisorSupportMatrix
Public vs. private, Amazon compared to OpenStack. http://redhatstackblog.redhat.com/2015/05/13/public-vs-private-amazon-compared-to-openstack/
Mirantis comparative cost calculator for both AWS and OpenStack. https://www.mirantis.com/comparing-openstack-and-aws-total-cost-of-ownership/
OpenStack Architecture Design. https://docs.openstack.org/arch-design/_images/osog_0001.png
The squirrelbin microservice architecture. https://aws.amazon.com/blogs/compute/the-squirrelbin-architecture-a-serverless-microservice-using-aws-lambda/