Segmentation and Replication of a Database

Segmentation Schemes of Databases

Guru Prakash Reddy Thuthipattu

MSc. in Data Analytics – Full Time

National College of Ireland

Dublin, Ireland

X16135369@student.ncirl.ie

Abstract—Segmentation and Replication of a database is one of the major factors that affects the performance of a data warehouse and in this paper we explore the different approaches of how it is done and compare it. The main objective this study is to determine which will be a better segmentation scheme in comparison.

Keywords—database; segmentation; fragmentation; cluster; replication; datawarehouse;  

1.  Introduction

In this age, where businesses are highly dependent on data for the running of the business it is essential for us to construct a data ware house that not only incorporates huge volumes of data collected but also give us a response for a query quickly,

One of the major factors that governs the performance of a data warehouse is how the databases that are linked to it are structured. We can have a single database which has all the transaction reports or we can fragment the database based on some characteristic features like demographic profile. In the first type, when we search for a query it must look at all the contents (rows) until we get the desired details. But in the latter case, as we are fragmenting the database we will be looking up at a segment of the main database which has similar profile as our search query. Fragmenting of the database helps us to reduce the time response for fulfilling a query.

1. Fragmentation And Replication

1. Need for Fragmentation

It becomes very difficult and time consuming for us to get a response if we have a single database as we have to run through every single record. It becomes easier for us if we are able to look into a more customized database where the results for our search string might lie. This can be done when we divide the database basing on some characteristics and reducing its volume which is tailored for our needs.

1. Process of Fragmemtation

According to Ajit M.T and Sudha Ram (1998) [1], the process of fragmentation can be divided into the following steps

• Fragmentation of the database into multiple pieces without the loss of any data
• Allocation of locations to the segmented databases
• Replication of the segmented databases which helps us improve the system performance. We should be very careful with the replication process as having more or lesser copies will not give the desired performance.
• The replicated fragments should be synchronized and must be consistent.
• We need to take into account that a query might require data from multiple segments and optimize it.

According to Abraham Silberschatz, F. Henry Koth and S. Sudarshan (2009) [2] , Fragmentation can be mainly classified as Horizontal and vertical

• Horizontal Segmentation: The database is divided into subsets with reference to one or more key attributes
• Vertical Segmentation: Here the subset is generated by “dividing the attributes of a relation into groups and then projecting the relation over each group.

1. Replication

According to Amit of data followed by Secondary Distribution of data. In the first step we look at the information that is required for making key decisions and we us this data for construction of the distribution matrix. In the second part we look into the transactional part of the database which is based on the distribution matrix developed.

1. Segementation Schemes

The segmentation scheme adopted by Ajit M.T and Sudha Ram (1998) [1], involves Primary Distribution of data followed by Secondary Distribution of data. In the first step we look at the information that is required for making key decisions and we us this data for construction of the distribution matrix. In the second part we look into the transactional part of the database which is based on the distribution matrix developed.

1. Adaptive Segmentation

Milena Ivanova, L. Martin Kersten  and Niels Nes (2008) [3], have proposed a segmentation procedure which will take into consideration the queries of the  past that will be useful in the future. They have developed it in a way so that it will be able to reorganise the data without any further help.

1. Adaptive Segmentation used in [3], can be done in two ways, randomized decision making and adaptive pagination

2. RandomiEquations

The equations are an exception to the prescribed specifications of this template. You will need to determine whether or not your equation should be typed using either the Times New Roman or the Symbol font (please no other font). To create multileveled equations, it may be necessary to treat the equation as a graphic and insert it into the text after your paper is styled.

Number equations consecutively. Equation numbers, within parentheses, are to position flush right, as in (1), using a right tab stop. To make your equations more compact, you may use the solidus ( / ), the exp function, or appropriate exponents. Italicize Roman symbols for quantities and variables, but not Greek symbols. Use a long dash rather than a hyphen for a minus sign. Punctuate equations with comm,as or periods when they are part of a sentence, as in