Annotated Bibliography Business Intelligence

Question :

Explain how Database system serving needs of Business intelligence

Answer :

In this process regular information is not much effective or useful. The platforms of Business intelligence are a highly developed and advanved in terms of technology and can be used as a tool to test any innovations in advanced business informing (p. 2). Development in technology helps in developing certain level of expectations from the end users. In addition to this, there are wide range of problems arrises implementing business intelligence, developing expected value of the business between effective implementation of business intelligence and advanced technology comes up. There are various reasons for these issues the expectations of value are relatively higher while making any investment in a sophisticated information system such as business intelligence system. Although, There are various types of database systems which can be used in an organisation to enhance performance, Relational, Hierarchical and Network database systems. 

• Relational database model is type of database that stores and provides access to data points that are related to one another. There are various types of database systems such as Microsoft SQL, Database as a service etc. 

• A Hierarchical  database model is a data model in which the data are well organized into a tree like structure. IBM and RDM are the most popular Hierarchical database systems used in organisations. 

• The last is network database system is a database model conceived as a flexible method of representating objects as well as their relationships fir example, IDS (Integrated data store) is a most used network database system.

Digital transformation has a great potential of changing lives of consumers and creating business value. Improved service quality, Enriched decision making, enhanced revenue and speed to market are some of the major advantages contributing since the inauguration of big data implementation across organisations all over the world (p. 1). In last decades, Fourth industrial revolution has determined the huge increment in availability of data or information, decision makers in both private and public sectors have identified new uncertainties related to future of operations or production. Under the right situations, the decision makers will be able to capitalise and manage on imminent transformation. Big data is applicable to information that can't be analysed and processed utilising traditional tools or procedures. Many businesses today are facing great issues although they have obtained information but it is unfortunate for not able to utilise those attained data or information fully. Big data analytics is the most studied topic in the field of Business Intelligence, this is because the core data analytics and data mining was determined during the early phase of business intelligence framework. The major objective of Business Intelligence model is providing accurate insight with relevant data or information, in order to assist a company to attain comprehensive information influencing the business or its operations. Business intelligence is an interactive computer based systems or structures with subsystems that helps decision makers to use data, knowledge, documents, predictive and analytical models easily. It further assists in identifying and justifying the issues in a tangible manner (p. 2). By using business intelligence application organisation can improve strategic and operational performance. However, Business intelligence technologies are providing current, historical and predictive views of business related activities. The organisational credibility of business intelligence technologies provides analytical reporting from online analytical processing (OLAP) (p. 2). On the other hand, more complex data mining, process mining, analytics, complex event processes, benchmarking, business performance management, predictive analytics, text mining, machine learning and prescriptive analytics as well as deep learnings can be attained from data science space. 

Although, from the database research perspective, one question arises in this circumstance, why these systems are more effective than classical distributed database, they have been used by many organisations already. The answer to this question lies in Byzantine fault tolerance, while classical distributed database systems need a trusted set of participants(p. 105), Blockchain system are capable of dealing with a particular level of maliciously behaving nodes. This attribute opens a wide range of application fields for instance, transactions between two organisations, that do not trust each other. With respect to the aspect of Byzantine fault tolerance, blockchain systems are much more effective as compared to distributed database systems. Unfortunately, any other element of transaction processing, classical database system are definitely ahead of Blockchain systems. A great example for this is the order execute transaction processing model, that prominent systems such as Ethereum and Bitcoin implement. In the stage of ordering, all managers first agree on a global transaction order, specifically utilising a consensus mechanism. Then, each entrepreneur is executing the transaction locally in that order on a similar state. While this mechanism or approach is simple, it has two major downfalls (p. 105). Firstly, the execution of transactions occurs in a sequential fashion. Secondly, as each and every transaction needs to be executed on every business owner or leader, the performance of the system does not scale with the numerous peers. On the other hand, there are blockchain systems that are trying to be more effective and beneficial. Hyperledger Fabric is a best example to this, a highly popular open source blockchain system offered by IBM. Instead of Implementing the order execute model, it is following a sophisticated simulate order validate commit model. 

It has been determined that strong similarities of the transaction pipeline of contemporary blockchain systems with respect to Distributed database and Hyperledger Fabric systems in general. 

There are a various implementations of the traditional relational model and triple stores along with the increasing NoSQL technologies (p. 1). The study determined that for a develop who require to select one database system, it is very essential to understand how the data stored by an application looks like and what kinds of use-cases the system will be addressed with for ensuring the system is appropriate for reaching optimal overall query performance. Relational databases have been introduced since many decades, it in a most leading choice for the developers of various technologies regarding making decision of how to store the information or data of an application (p. 2). In recent years, developers prefer moving towards technologies outside the relational model, that incorporates different ways and solutions of storing relevant data. The main assumption of this study is that no individual developer can identify which is the best available database system in advance, best for queries of an application and for the data. If the decision making is effective and appropriate, an analysis is conducted each and every time an application is designed and developed (p. 2). 

It tends to have various structural attributes. For example, there are areas with huge number of tree like struo and many-to-many relationships, which are likely to be visible in graph based models (p. 4). Those models can be seen in applications requiring high interconnectedness of their types of entity such as social media platforms. Various other models can also be identified which are storing string or numerical based data in a simple tabular format for instance, a financial application or address book. Each and every database system uses a different type of internal data structure dealing with the data it stores. A data model that is developed of attributes that are typical for graph structures, a graph database might be best suitable (p. 5). Although, each and every query is addressing diverse parts of a data model within the database, a developer may not be able to determine which is database is most appropriate for an application in advance. A best suitable database system is one that may serve with the highest performance in each and every query of an application as compared to any other database systems. For the purpose of predicting the runtime is several queries addressing a schema, an algorithm by means of a cost function has to be created, it considers a wide range of characteristics or features. The algorithm tends to evaluate the value which further represents a level of suitability of the provided range of queries addressing the siven schema, to a list of a specific set of database systems of divers forms. It is decided to utilise the technique from the field of machine learning for solving that particular issue (p. 6). 

References

Zloch, M., Hienert, D., & Conrad, S. (2017). Towards a Use Case Driven Evaluation of Database Systems for RDF Data Storage-A Case Study for Statistical Data. In BLINK/NLIWoD3@ ISWC.