An Analysis of the NoSQL Database Management System - Essay Example

An analysis of the NoSQL database management system and Introduction With the recent advancement in technology, enormous data is generated and transferred through the web. Through applications such as Facebook, Twitter, Google among others, access to and use of data is becoming easier. With social graphs, personal user data, machine logging and content generated by users have caused an exponential increase to data. For such services to be availed properly, huge amounts of data are processed - a service that traditional database management system could not handle (Pludge & Membrey, 2010). A critical analysis of the NoSQL database management system shows that it can be a perfect solution for the exponentially growing data volumes.It employs distributed computing approach and would be a perfect choice for managing such big data. About NoSQL The term NoSQL has been denoted as “not only SQL” by many people. It however has no any universally accepted definition. This title is affiliated to the fact that NoSQL databases do not require execution of SQL statements to query the database. It was devised by Carlo Strozzi in 1998 while referring to his Light Weight, Open Source Database that had no any SQL interface.It is designed in a manner that data storage does not take the form of schemas, thereby avoiding join operations and making it easy to scale horizontally. In 2009, Eric Evans from Rackspace Inc. also used the term while referring to databases which are distributed and non-relational, and which did not conform to the ACID properties of the traditional relational database management systems. Later discussions and use of NoSQL got an impetus that saw it growing remarkably. NoSQL is a group of eclectic non-relational data management systems in which the databases does not use the table schemas to store data. Manipulation of these databases does not rely on SQL as is the case of the conventional relational database management systems. NoSQL databases have gained much impetus due to their ability to support huge quantities of increasing data. With the application of the distributed systems concept, NoSQL databases are designed to handle massive data processed parallel across many servers, employing the use of non-SQL mechanisms and languages to manipulate data. This is achieved through various APIs that aid in converting SQL statements into the system’s innate query language. They emerged as a solution to the challenge major internet companies were facing while dealing with huge quantities of data. Characteristics of NoSQL To guarantee data integrity, most conventional database management systems employ the aspect of transactions. This is in a bid to ascertain data consistency in all aspects of database management. It constitutes to what is referred to as the ACID properties for database management common among the relational database management systems. The ACID properties include Atomicity, Consistency, Isolation and Durability. To scale out of the ACID compliance has for long been perceived as a problem. Issues have arisen between the varied aspects of high data availability in distributed systems that could not be solved in totality, and, therefore, leading to what is known as the CAP theorem. This theorem works on the basis that when scheming applications for distributed systems, three requirements exists which should be met. These include Consistency, Availability and Partition-Tolerance. Consistency: This principle requires that all clients should have access to the same version of data, even after performing an operation to the database. For instance, all the users of a system should see similar data after an update to the database. Availability: the system should always be on, thereby guaranteeing service to the users. Partition Tolerance: this property requires that the system should remain functional even when the communication among the servers becomes unreliable. It should give room for the servers to be partitioned into different units that do not necessarily need to communicate. Fulfilling all these requirements in totality is impractical. The CAP theorem by Eric Brewer provides for a distributed system to achieve any two of the three aspects. Many of these databases have however slackened up on consistency in a bid to achieve high availability and partitioning. They thus fall under a category of systems with the BASE architecture. The acronym BASE, stands for Basically Available, Soft State, Eventually Consistent systems. Fig.1 Characteristics of NoSQL Database (Source:nosqltips.blogspot.com ) In regards to the CAP theorem, Basically Available implies that the availability of the system is not guaranteed. The Soft state on the other hand requires that the system should be prone to changing over time, even without any alterations. Finally, the Eventual consistency aspect requires that the system will be consistent over time even without any inputs during such time. The BASE systems lack transactions in the conventional nous and presents constraints on the models of data, and thus enabling partition structures (Han, Haihong, Le, & Du, 2011). Primarily the NoSQL databases are used for processing large-scale data, embedded information search and retrieval, large-scale data storage and exploratory analytics. These database management systems employ a distributed framework which makes them an ideal choice for massive batch data crunching and processing. Classification of the NoSQL databases According to Leavitt (2010), NoSQL databases are broadly grouped into three categories. These include Key-Value stores, Document databases and Wide-Column stores. There are however other categories of NoSQL databases that exist which includes the Graph databases and XML databases based on the XML format. The Document oriented databases stores data in the form of documents. This is common in databases such as the MongoDB, RavenDB and the CouchDB. XML databases stores data in XML format and constitutes such databases as BaseX and MarkLogic. Graph databases store Data in nodes that are analogous to objects in a programming language. These nodes are connected to one another by edges. Neo4j, Allegro and OrientDB are examples of Graph databases. The Key Value stores employ a schema-less architecture in information storage. A key could be lists, hashes, sets or even strings in which values are stored against. Dynamo, MemcacheDB and Redis fall under this categorization. Finally, the Column store category which includes HBase and the Facebook’s Cassandra employs a key value pair in which the key acts as an identifier and the value as a store for related data. NoSQL pros and cons Advantages NoSQL exhibits elastic scaling in which the database is distributed across manifold hosts with an increase in data load. This feature is not possible with the conventional database management systems which rely on scale up, in which bigger data servers are acquired with a load increase. This ends up being more expensive and thus an advantage with the non-relational database systems. With the increases in the rates of transactions over the recent past, the volumes of stored data have also increased massively. The capacities of RDBMS has also been growing to equal the upsurges. However, the rates of transactions and the constraints resulting from data volumes that can only be managed by one RDBMS are getting intolerable for enterprises. This has however been solved by the presence of NoSQL systems as they can handle very huge data that even the biggest RDBMS cannot. NoSQL database management systems have been designed and build in a manner that makes them require less management. Organizations thus require not to higher the expensive highly trained experts to manage. They are designed with automatic repair features and data distribution, thus becoming economical and efficient for most enterprises. The use of cheap servers in managing the big volumes of gradually expanding data makes it easier for enterprises to manage large volumes of data at a cheaper price. Finally, NoSQL lacks data model restrictions. The NoSQL Key Value data stores and the document data stores give room for applications to store almost any edifices available in the space provided. With this feature, the databases allow new columns to be generated without too much hassle. Disadvantages Despite the many promises associated with NoSQL, there are a number of obstacles that should be overcome first for them to remain appealing to the mainstream enterprises. Among such challenges includes maturity. The maturity of most RDBMS according to most database professionals is reassuring. They are said to be stable and highly reliable. However, for NoSQL, many alternatives are still in their preproduction stages with most of the prominent features yet to be implemented. Lack of support staff and administration professionals also makes these systems fail to capture the interests of major enterprises. Conclusion The use of NoSQL database approach is gaining status with a gigantic speed. However, this does not imply that the conventional relational databases are becoming undeveloped or something ancient. Their use will most likely remain active, but not in isolation with the new and popular high performing NoSQL databases. Rather, the two will be used closely in what could be termed a symbiotic association. The current age could be termed as an era of polyglot tenacity, in which different data needs are met using data from different warehouses. In such a time, the monopoly that relational databases used to enjoy can never be enough as there will always be need for alternative data. Gradually, more complex system are required by different architects which will meet their needs and requirements based the nature of data to be stored and the volume of such data. This should therefore act in favor of such individuals and should allow the highest level of flexibility in storage and manipulation of data. References Han, J., Haihong, E., Le, G., & Du, J. (2011, October).Survey on nosql database.In Pervasive Computing and Applications (ICPCA), 2011 6thInternational Conference (pp. 363-366). IEEE. Plugge, E., & Membrey, P. (2010).The definitive guide to MongoDB the NoSQL database for cloud and desktop computing. New York, NY: Apress. Tesoriero, C. (2013). Getting started with OrientDB: a practical guide to learn, deploy, and customize OrientDB. Birmingham: Packt. Read More