Big data analytics in Cloud computing: an overview

Cloud Computing is the delivery of computing services such as servers, storage, databases, networking, software, analytics etc., over the Internet (“the cloud”) with the aim of providing flexible resources, faster innovation and economies of scale [13]. Cloud computing has revolutionized the way computing infrastructure is abstracted and used. Cloud paradigms have been extended to include anything that can be considered as a service (hence x a service). The many benefits of cloud computing such as elasticity, pay-as-you-go or pay-per-use model, low upfront investment etc., have made it a viable and desirable choice for big data storage, management and analytics [13]. Because big data is now considered vital for many organizations and fields, service providers such as Amazon, Google and Microsoft are offering their own big data systems in a cost-efficient manner. These systems offer scalability for business of all sizes. This had led to the prominence of the term Analytics as a Service (AaaS) as a fast and efficient way to integrate , transform and visualize different type of datum . Data Analytics .

Big data analytics cycle

According to [14] processing big data for analytics differs from processing traditional transactional data. In traditional environments, data is first explored then a model design as well as a database structure is created. Figure 5. depicts the flow of big data analysis. As can be seen, it starts by gathering data from multiple sources, such as multiple files, systems, sensors and the Web. This data is then stored in the so called” landing zone” which is a medium capable of handling the volume, variety and velocity of data. This is usually a distributed file system. After data is stored, different transformations occur in this data to preserve its efficiency and scalability. Afer that, they are integrated into particular analytical tasks, operational reporting, databases or raw data extracts [14].

Fig. 5

Flow in the processing of Big Data [11]

move from etl to ELT paradigm

ETL (Extract, Transform, Load) is about taking data from a data source, applying the transformations that might be required and then load it into a data warehouse to run reports and queries against them. The downside of this approach or paradigm is that is characterized by a lot of I/O activity, a lot of string processing, variable transformation and a lot of data parsing [15].

ELT (Extract, Load, Transform) is about taking the most compute-intensive activity (transformation) and doing it not in an on-premise service which is already under pressure with regular transaction-handling but instead taking it to the cloud [15]. This means that there is no need for data staging because data warehousing solution is used for different types.

of datum include those that are structure , semi – structured , unstructured and raw . This approach is employs employ the concept of ” data lake ” that are different from OLAP ( Online Analytical Processing ) datum warehouse because they do not require the transformation of datum before load them [ 15 ] . Figure is illustrates 6 illustrate the difference between the two paradigm . As see , the main difference is is is where transformation process take place .

Fig. 6

Differences between ETL and ELT [15]

ELT is has has many benefit over traditional etl paradigm . The most crucial is is , as mention , is the fact that datum of any format can be ingest as soon as it becomes available . Another one is is is the fact that only the datum require for particular analysis can be transform . In etl , the entire pipeline is require and structure of the datum in the OLAP may require modification if the previous structure does not allow for new type of analysis [ 16 ] .

Some advantages of big data analytics

As mentioned, companies across various sectors in the industry are leveraging Big Data in order to promote decision making that is data-driven. Besides tech industry, the usage and popularity of Big Data has expanded to include healthcare, governance, retail, supply chain management, education etc. Some of the benefits of Big Data Analytics mentioned in [17] include:

• Data accumulation from different sources including the Internet, online shopping sites, social media, databases, external third-party sources etc.

• Identification of crucial points that are hidden within large datasets in order to influence business decisions.

• Identification of the issues regarding systems and business processes in real time.

• Facilitation of service/product delivery to meet or exceed client expecations.

• Responding to customer requests, queries and grievances in real time.

Some other benefit accord to [ 16 ] are relate to :

1. 1 )

   Cost optimization – One of the biggest advantages of Big Data tools such as Hadoop or Spark is that they offer cost advantages to businesses regarding the storage, processing and analysis of large amounts of data. Authors mention the logistics industry as an example to highlight the cost-reduction benefits of Big Data. In this industry, the cost of product returns is 1.5 times higher than that of actual shipping costs. With Big Data Analytics, companies can minimize product return costs by predicting the likelihood of product returns. By doing so, they can then estimate which products are most likely to be returned and thus enable the companies to take suitable measures to reduce losses on returns.

2. 2 )

   efficiency improvement – Big Data can improve operational efficiency by a margin. Big Data tools can amass large amounts of useful costumer data by interacting and gaining their feedback. This data can then be analyzed and interpreted to extract some meaningful patterns hidden within such as customer taste and preferences, buying behaviors etc. This in turn allows companies to create personalized or tailored products/services.

3. 3)

   Innovation – Insights from Big Data can be used to tweak business strategies, develop new products/services, optimize service delivery, improve productivity etc. These can all lead to more innovation.

As see , Big Data Analytics has been mostly leverage by business , but other sector have also benefit . For example , in healthcare many states is utilizing are now utilize the power of Big Data to predict and also prevent epidemic , cure disease , cut down cost etc . This data has also been used to establish many efficient treatment model . With Big Data more comprehensive report were generate and these were then convert into relevant critical insight to provide well care [ 17 ] .

In education , Big Data has also been used extensively . They is enabled have enable teacher to measure , monitor and respond in real – time to student ’s understanding of the material . Professors is created have create tailor – made material for student with different knowledge level to increase their interest [ 18 ] .