Methodology for Choosing Data Centre Location in West Africa Region

Methodology For Choosing Data Centre Location in West Africa Region

Ahmed Alharthi’s Master Project Proposal

Last Update: Thursday, March 5, 2014

Committee Members and Signatures:

1.  Introduction

This project proposal report gives a description of the research and process which will be undertaken in delivering a methodology for selecting the best location to build Data center in West Africa. Different factors will be study and analyze when considering the best location for the Company to base its new data center away from its regular locations [1]. The project will mainly analyze a country in West Africa within which the most Companies do not have a single data center. The final research paper will consist of report document and a web-based application google map that will be used to demonstrate the collected data.

There are many studies, reports and expert demonstrations on what are the most important elements in deciding the best location to build a data center. There are different factors considered when answering this question including facilities, the efficiency of operations of the center and cost factor which among the three is the most important element and which is the most critical business decision for building and maintaining the infrastructure required. IT businesses, small and large are similar in the way they make decisions. In order to make the best decision for their company researchers and management require sufficient information before arriving at a decision and that is the purpose of this project. In this project, we attempt to go through the decision criteria and options available for IT businesses which are looking to build their data centers in West Africa. This project will attempt to produce a detailed list of the most important factors [2], [3], [4].

In this project, we will discuss the various risk factors that may impact data centers especially the geography in West Africa. Similar cases, such as the cost and geographic advantages of choosing Colorado for the Company’s data center operations will be analyzed by studying probabilities of natural disasters such as floods, tornados and hurricanes. There are other issues that should be considered in this project, such as the political situation in these countries. All these factors will be considered during this research and more issues may be uncounted as the research progresses.

1.1. Amazon EC2 Global Infrastructure

Amazon EC2 has nine regions and forty two edges around the globe, but it can be seen clearly from figure 1 that there are no data center facilities on African, and that is because of the lack of data availability on these regions. As result, Africa

Figure 1. Amazon EC2 global infrastructure [1]

Has fewer options to communicate efficiently with the Amazon infrastructure. The mentioned regions have more latency and higher power costs consumption when it comes to interact with Amazon EC2 environments. In this research we intend to suggest new data center locations for Amazon EC2 in West Africa.

1.2. Related Work

On related work, the paper will analyze the presence of any previous work undertaken concerning the suitability of locating data centers. For example, John Rath has presented a work that attempt to breaks down the decision criteria and options available to companies who are looking to build data center. His work presents an exhaustive list of potential factors to be weighed when doing site selection analysis. [2]

IBM has well explained documents in data center strategies. For example, how to meet a growing need for more data center capacity, the IBM Canada Leadership Data Centre leverages best practices developed in designing, building and operating data centers around the world. [5]

Other related work, Federico Larumbe, Brunilde Sans`o have presented work on Cloud Location and Routing problem (CLRP) [6]. A mathematical problem that is looking to compute geographical distribution on multiple rejoins, the location of data center, servers and software components. Their results underline the importance of location and routing showing that the budget and number of datacenters have critical impact on overall network performance.

Khosrow Ebrahimi, Gerard F. Jones, Amy S. Fleischer have presented a comprehensive review of data center cooling technology, operations conditions and corresponding availability of data center waste heat sources and stream.[7]

Linda K. Nozick, Mark A. Turnquist has presented a model which forces to answer these questions, how many Data center should be in a place? Where should be located? Also, the p-dispersion problem is closely related to this proposal. The p-dispersion problem seeks to locate p facilities in an area or a graph such that the minimal distance between two facilities is maximized [8].

There are other public researches about strategies of picking up data center locations for Africa. In same time, there are reports about the most important factors for choosing data center locations in United States for army and enterprise [2], [4]. Also, there are different electronic journals that talk about different factors about preventing specific locations [3].

Other data will come in handy since it gives a preview of what has already been researched and what needs to be researched further. Critical issues and factors which need to be analyzed when determining the location of data centers will also be researched, by evaluating work done on factors to consider when deciding the best location for data centers. Research on related work will span different regions, factors to consider, mitigating risks and also methods of undertaking such kinds of research.

1.3. Proposed Project

This project will provide a methodology to suggest the best location of setting up data centre in a country in West Africa region. The methodology will consider important factors including cost, cloud performance observed by the users, network capability, environmental factors, economic stability, politics system. Network capability and setting up the data centre cost are two of the most important matrices of this project [24]. By studying these matrices I should be able to develop a methodology that helps to determine the best location for setting up the data center for other country [9]. Google maps come in handy to find and demonstrate the best location for setting up datacenter. This will be an extensive approach since the project initially has to look at all West African countries. The proposed methodology should help to determine the best location for to setting up a data center in other countries. The results should help organizations or companies interested in setting up data centers in West Africa countries. It will help their users interact much better with data centers provided by the cloud providers with low latency and high throughput.

2.  Significance of the Research

This research is important in the current age of technological advancement. Currently globally, IT is a major factor that boosts the economy of most countries. The demand for fast and reliable internet connectivity is growing. This demand helps in connecting various countries and making the world to be a global village.

West Africa is a growing IT hub and it is no exception for it to be connected to the rest of the world. Currently data providers are limited in this region. For this reason, the economy is boosted mostly by agriculture and industries. This therefore creates a good environment for setting up a new data center. This is because the new data centers are assured of a major market share in the region because they improve delivering services for customers. This in turn will help to create more IT related businesses. At the same time the rolling out of new internet connectivity will help dispense information [10].

This will help to boost other related sectors of the economy. Education will be greatly improved as stakeholders will be able to carry out more innovative ideas. The information and technology sector, mostly media houses, will be able to improve. They will be able to create more digital content and at the same time create more job opportunities [11]

Amazon EC2 can therefore look at the wide opportunity to invest and expand their presence. At the same time, it will be able to create more revenue and profits. At the end, this research will enable the economy to grow. Businesses that depend on data centers are many and this will surely affect everybody in the region. For this reason West Africa will be the new investment hub that investors won’t ignore [12].

3.  Problem description

There is lack of sufficient research and information about Africa regarding possible best locations in which data centers can be put up to aid in the globalization movement that has led to ease of doing business and world trade [13]. For this reason, this research paper will focus on West Africa region and analyze and evaluate one of the following countries with an aim of proposing the best location for Amazon EC 2 as an example to base its data center within these geographical regions:

Senegal, Guinea, Liberia, Ghana, Togo, Benin, Sierra Leone, Nigeria, Cameron, Cote d'ivoire, Congo, Angola, Democratic Republic of the Congo and Namibia.

These countries have been chosen since they lie next to the large water bodies and these oceans and seas are serviced by marine cables such as fiber optic which are a major internet and communication infrastructure [14]. Furthermore, Amazon EC2 has requested an in depth evaluation of these countries since it plans on having a data center in these parts of the world to help the company with its service delivery and expansion plans into new markets and regions.

3.1. Targeted audience

The targeted audience for this research is any one conducting research in cloud computing and the related facility location problem, in particular to those cloud providers interesting in set up data centers in West Africa. For example, Amazon are planning to expand their operations into new regions and based on their current market presence as shown in Figure 1 Company’s global infrastructure. The Company currently does not have data centers in Africa. Any one intends to do that needs a critical evaluation of geographical, risk mitigation, natural disaster factors and man-made situation factors within the listed countries in order to determine the best location for their new data center. Although the paper will give a clear and precise location of the most suitable location based on the proposed metrics, the final decision on where the data center will be located will depend on other factors such as demands and financial situations.

4.  Research Plan

The research will involve analyzing the considered factors in section 4.1. These factors will be evaluated critically for a location to be chosen as main location for a cloud provider’s next data center outside its current locations [15].

The audience will be taken into account. This is to enable them be convinced with the research. This is to show that it is a high impact project so as build investor confidence on the location that will be selected. This should include detailed cost estimation [16].

4.1. Considered factors

This project is trying to deal with the following factors in order to reach to the best location on a country:

a.  Cost of data center includes

§  Buildings.

§  Electricity (yearly expense).

§  Landscape [17].

b.  Data center distance from submarine cable [18].

c.  Network capability includes

§  Network latency (RTT)

§  Throughput

§  Average Typical Packet Loss [19], [20], [21].

d.  Environmental factors [2], [22].

e.  Economic stability.

f.  Politics System.

g.  Ease of can access to the suggested location.

4.2. Schedule and Tasks

- Research done during fall 2013

1. Research economical, costs and characteristics of different countries.

2. Research the basic needs of establishing data center facilities.

3. Research the most reliable marine cables such as, South Atlantic Express (SAEx), Atlantis-2, SAFE, South Atlantic Cable System (SACS), SAT-3/WASC, SEACOM/Tata TGN-Eurasia, FLAG Europe-Asia (FEA), The East African Marine System (TEAMS) and others more.

4. Research related information for marine cable such as, cable length, owners, landing points and design capacity.

5. Estimating and calculation distances between different data centers and edges.

6. Read about common data center facilities.

- In Progress - Work during spring 2014

1. Research on different locations, facilities and edges.

2. Collect existing information, facts, and considerations about picking up locations.

3. Research the most important consideration factors for choosing locations.

4. Perform simulation for choosing data center locations for Amazon EC2.

5. Writing the final report the process of this research with well detailed explanations, facts, implementations, experiments, and so on.

6. Design google map to demonstrate location’s data.

5.  Methodology and Technology

When deciding the location for the project numerous methods of data collection and analysis will be considered [2], [5]. Data will be collected from different publications, government departments and agencies that had already collected data. These data will be weighed and then calculated to give measurable rating to decide whether or not a given location is good to locate data center facilities [7], [4]. The gathered data will be compared with result from different countries to ensure consistency. It this will be important so as to select the best suitable location for setting up the data center [23]. Figure 2 shows data flow of decision criteria of this project

Figure 2 data flow diagram of decision criteria

6.  Deliverables

The deliverables for this project will include a written document and web-based application. The first document will include the results of research of the chosen country. This will include the evolution of the all the previously mentioned factors in section 4.1 which have been set out as the main factors to consider when deciding the best location to set up the data center. A program that will demonstrate all the data for the chosen the country using google map’s API as Figure 2 shows.