Relational Database Management System Implemented by Amazon Inc.

Relational Database Management System Implemented by Amazon Inc.

Organization Background

Amazon Inc. Goals and Objectives

The company is guided through four main principles that include competitor focus, customer obsession, commitment to operational excellence, and long-term thinking (About Amazon Staff). The company’s overall goal is to be earth’s most customer-centric company. The idea is to build a platform where customers can find what they are looking for and further discover other aspects while purchasing online (About Amazon Staff). To achieve this, Amazon Inc.’s mission is to offer consumers the lowest possible prices, improve selection varieties, and at the utmost convenience (Amazon.com).

Aim and Objectives of the System

Amazon Inc. uses a relational database management system as its core information system. RDBMS is regarded as a collection of capabilities alongside programs that allow information technology specialists and stakeholders to create, update, administer, and interact with a relational database. Amazon relational DBMS is a web service that permits users to make setups, operations, and scalability options in relation to the AWS cloud (Amazon Relational Database Service User Guide). Therefore, a relational database utilizes a structure that permits the users to identify and access information about another piece of data within the same database (AWS.Amazon). An RDBMS stores data in a tabular form, allowing the connection between relatable data items.

A relational database management system aims to store information. An RDBMS allows the users to generate a set of relational schema, which enables them to store data without unnecessary redundancy and retrieve information quickly, as shown in figure 1. The primary objective of an RDBMS is to provide an integrated collection of relatable files along with their interpretation. Therefore, the DBMS allows the users to retrieve or access the data within the database. Moreover, relational-DBMS allows users to manipulate data contained within the system in a more convenient, non-technical, and effective manner. Relational-DBMS has an added objective of storing data in a centralized system (AWS.Amazon). The advantage of such structures is that they prevent unauthorized individuals from accessing the information, therefore, ensures the privacy of

Figure 1: An example of data within an RDBMS. Source:

Stakeholders of Oracle RDBMS

The first stakeholder is the portfolio management team. This category could also include a software factory and service competency center (Amazon Relational Database Service User Guide). The roles and responsibilities of this stakeholder are to identify the standard needs across the users, determine the assets and levels of investments required, and expected return on investments (Amazon Relational Database Service User Guide). Furthermore, the stakeholders can establish a workflow portfolio, participate in the administration of the database, and evaluate the performance of each packet within each database.

System Development Methodology

Amazon Inc. utilizes a sharded database architecture. Sharded is a term that refers to the splitting up of information or data into smaller units with a consequential distribution across a number of platforms such as servers (AWS). Therefore, each server is known as a shard. In Amazon, the servers are in multiples, but the unique aspect of it is that they share common features. Among these include the same engines, data structures, and type of hardware (Amazon Relational Database Service User Guide). However, each server has no knowledge about the existence of another server, which permits scalability in approaches such as clustering or duplication (AMS). When developing the relational DBMS, Amazon’s concept includes a share-nothing model within the sharded database. The architecture safeguards from fault tolerance and scalability options (Amazon Relational Database Service User Guide).

Current Software Capabilities

Hardware

Through the DB instances, Amazon’s physical storage is considered in numerous platforms. They include MariaDB, Oracle, PostgreSQL, Microsoft SQL Server, and MySQL (Amazon Relational Database Service User Guide 29). There are three types of relational DBMS storage types, and they include General purpose SSD, provisioned IOPS, and magnetic (Amazon Relational Database Service User Guide 29). The general purpose is regarded as one of the most cost-effective physical storage components that is ideal for large workloads. The aspect of the task is through the assignment of single-digit latencies, which boosts about 3 000 IOPS for extended periods of time. The ability to do so permits baseline performance, which in translation is volume to volume size workloads. The other is provisioned IOPS. It is storage designed to meet the needs of the I/O workloads. They are particular to database workloads, which require limited I/O competencies and consistency in throughput (Amazon Relational Database Service User Guide 29). The last is magnetic, which supports any magnetic storage for backward compatibility. According to Amazon Relational Database Service User Guide, it is recommended that general-purpose SSD and provisioned IOPS to work hand in hand with magnetic storage. The reason is that the amount of storage permitted for magnetic is less compared to provisioned IOPS and general-purpose SSD (Amazon Relational Database Service User Guide 29).

Software

The software composition of Amazon RDBMS is provided through the cloud computing framework. Currently, Amazon provides state of the art technology with regard to data centers. Data is specific to given locations, which are composed of AWS regions and availability zones (Amazon Relational Database Service User Guide 34). The users are provided with the ability to access data regardless of the region based on the resources required, even if it is in multiple locations. However, the Amazon Relational Database Service User Guide indicates that there could be failures experienced regarding the use of the availability of resources within the same location (34). For instance, it is expected that if the information is located in one region or a single region, the possibility of failure is provided if there is no instance of availability. Therefore, it is recommended that each software component be independent. The reason is that for each Amazon relational DBMS activity initiated by the user, there is a need to ensure that it is the only operation working at the moment for each AWS region. Thus, the default is set to be changed within each console by setting the environment variable under the SQL function (Amazon Relational Database Service User Guide 41).

Another aspect regarding the software is the provision of support through the multi-AZ deployment for failover support for DB instances. There are several technologies that are in use, including Multi-AZ deployments such as Oracle, MariaDB, and Postgre SQL DB instances (Amazon Relational Database Service User Guide 41). The multi-AZ is an automatic provision under DB instances within each availability region. It provides synchronization of data across multiple regions, facilitating reduced redundancy and eliminates I/O freeze as well as minimize latency during system back-up (Amazon Relational Database Service User Guide 41).

Figure 2: A representation of relational-DBMS Oracle software components. Source:

Database

The database used by Amazon Inc. utilizes a pre-existing sharded design. In this, the data is segmented horizontally and distributed through partitions across a database shard (AWS). The user can have various strategies in partitioning the table, including the provision of range partitioning, hash partitioning, and list partitioning. The user is then permitted to use each database shard to accommodate more than one table partition with separate tables, as shown below. In case there are more than two types of tables involved, then a relationship key is used. To achieve this, AWS recommends the use of horizontal partitioning by using the same partition on all tables, as shown below.

Figure 3: A Representation of Database Source: https://aws.amazon.com/blogs/database/sharding-with-amazon-relational-database-service/

Telecommunications and Internet

The most significant diversification of business functions from DB Instnace is the integration of multiple internet and telecommunication applications. They range from search engines, social networking applications such as Facebook and cloud computing services (AMS). The processes utilize petabytes of data, which entail thousands of servers that serve millions of clients. The services include analytics and big data, storage, database management, networking, connectivity, and edge services (AMS).

Modeling Relevant Business Processes and Functions

Amazon Inc’s relational DBMS provides VMware web services. It is a business function that allows the user to deploy administered databases in on-premises that are housed by VMware environment using the relational database technology from which thousands of clients utilize the platform (Docs.aws). The primary aspect of the DBMS is to offer data concurrency and consistency. The business process in this concept is to allow the client to use data, modify it, or store it without being concerned that other users will modify the data at the same time. Preferably, the process is done under a multi-user database (Docs.AWS). Multiple transactions can be done and updated at the same time. Data concurrency ensures that users can access data at the same time. Concurrently, consistency provides users with a continuous view of their data, including visible data changes such as own transactions (Amazon Relational Database Service User Guide123). The unique ability of the relational-DBMS is guaranteed under a transaction isolation model known as serializability.

Another concept in business function is Amazon DynamoDB.  It is a key-value document database that delivers single-digit data at any performance scale (Docs.aws). Multi-versioning is a term under Oracle Relational-DBMS that allows multiple or simultaneous operations of numerous versions of data. The overall characteristic of multiple version data is that data integrity is guaranteed through read-consistent queries. According to the Business Database, the relational-DBMS permit dirty reads, whereby some transactions read uncommitted data in another operation. For example, Docs.AWS stipulates that more than 100 000 clients use the business platform for various aspects, including gaming, tech, IoT, and other applications. Therefore, relations databases offer business rules and policies. Under such business functions, strict rules about commitment exist. For instance, in an inventory database, some data versions are tracked, preferably in three parts. One part is placed under inventory while the rest is pulled. If one part is unavailable, the other two parts are pulled to make the database commit to a specific task.

Another business function is ElasticCache under Amazon Inc. It involves data access in repetitive functions through deployment, operation, and scalability of memory cache within the cloud. The relational –DBMS allows a single query to obtain information from a data table, which may need to be repeated multiple times or endlessly (Docs.AWS). Within the database, data access processes in any business processes require specific codes to access it. In most cases, application developers may not want to write new codes for the same functions in each novel application. Therefore, relational-DBMS permit storage of such functions in blocks, which can be accessed under simple application protocols. Therefore, a single stored process can provide for consistent tagging for users under multiple transactions.

Implementation Plan and Anticipated Difficulties in Implementation  

When implementing the Amazon Inc. relational-DBMS, several specific requirements are necessary. They include determining the memory capacity, CPU, and storage usage. The first process is to establish these requirements; for example, Amazon CloudWatch anticipates that the user needs to notify the system on the usage patterns or any change on the basis of deployment of data so as to notify the system for system upgrade (Amazon Relational Database Service User Guide123). The next process is to scale-up the DB instance when approaching storage capacity limits. The concurrent process was enabling automatic back-ups and set the back-up window to occur during the daily law in write IOPS (Amazon Relational Database Service User Guide123).  The following process entails using and viewing of performance metrics from this, the database design is implemented, and within the DM Instance. There are multiple designs used based on preference and objective ((Amazon Relational Database Service User Guide123). A logical schema specification is identified and then implemented. Mostly, the implementation process is embedded within the database, but a SQL function is necessary. The MySQL functions, for instance, entail the definition of data type for each column, including primary and foreign keys. Features such as constraints are required to promote the integrity of data in future use.

In the implementation process, several limitations are associated with relational database management systems. Primarily, a fixed schema presents problems. Amazon Inc. DB Instance has a pre-existing schema that may not necessarily adapt to every business or personal function. Therefore, whenever project managers or application administrators receive multiple requests to improve the experience in using the database, it becomes problematic and can take up a great deal of time (Trust Radius.Com). In addition, from a non-technical perspective, it becomes difficult to use multiple tables with related data. The outcome is that the complexity in computing resource increase; therefore, it becomes difficult to carry out a query at a given time. To resolve these issues, the relational-DBMS should allocate grid computing, which is a service that clients should have. A database server grid is a collection of commodity servers that together run on more than one database. Within these databases, one can build a pool of system resources which drastically allocate and de-allocate resources based on business priorities (Amazon Relational Database Service User Guide234). The added advantage of using the database grid structure is fault tolerance. Fault tolerance, according to a relational database, is the security provided by high availability architecture against any faults in the design or structure of the relational-DBMS in use.  Therefore, Amazon Inc. relational database has built-in integration for clients with respect to connection pools.

System Testing and Maintenance

System testing is conducted under specific performance through a dedicated core team in Amazon Inc. The purpose of the team is to improve and support customers who run on the database solution systems (Amazon Relational Database Service User Guide156). Among the testing, parameters include upgrading, networking, virtualization, performance, management, and experience. Other parameters include regression, conformance, and certification. For instance, with specificity to certification, the database testing team measures third-party certification, which entails assurance of clients about restrictions or permission policies regarding software use ((Amazon Relational Database Service User Guide156). Concurrently, the team performs maintenance on key features through upgrades. Information relayed to the administrators and project managers is pivotal to enhancing customer experience. Among the maintenance include storage, operating system, management, installment and deployment of software, upgrade of the I/O layout and performance, as well as injecting system faults within the I/O sub-systems so as to optimize operations (Amazon Relational Database Service User Guide126).

Training and Support

According to Houston, the company offers database training and certification. The objectives of the training and support are through tackling performance, availability, security, and compliance (Houston). The support given to clients entails understanding the modern features of the systems, comprehension of the performance, scale, and security features in addition to maintenance of the available databases. The training is specific to database architects, administrators, or data scientists (Houston). The training is done online or through a virtual class, either single-mode or through a classroom. There are options for training on demand under each package for the database management and administrative content.

Anticipated System Improvement and Future Plan

How to Use the System

Figure 4: Amazon Relational DBMS. Source:

Works Cited

Trust Radius. “Amazon Relational Database Reviews.” Accessed 23^rd March 2020 https://www.trustradius.com/products/amazon-relational-database-service/reviews

Houston, Sara. “Now Available: New Digital training to Help you Learn about AWS Database Services.” AWS. Amazon, 2018. https://aws.amazon.com/blogs/database/now-available-new-digital-training-to-help-you-learn-about-aws-database-services/

About Amazon Staff. “Our Leadership Principles.” Amazon. Accessed 23^rd March 2020 https://www.aboutamazon.com/working-at-amazon/our-leadership-principles

Docs. Aws. “Amazon Relational Database.” AWS. https://docs.aws.amazon.com/whitepapers/latest/aws-overview/database.html#amazon-rds

AWS. “Sharding with Amazon Relational Database Services.” Accessed 23^rd March 2020 https://aws.amazon.com/blogs/database/sharding-with-amazon-relational-database-service/

“Amazon Relational Database Service Guide.” 2020; pp 1-400. https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/rds-ug.pdf

AWS.Amazon. “What is a Relational Database Management System.” Accessed 23^rd March 2020 https://aws.amazon.com/relational-database/

Amazon RDS. “Amazon Relational Database Service.” Accessed 23^rd March 2020 https://aws.amazon.com/rds/