Management Of Physical Storage For Database Systems

Management Of Physical Storage For Database Systems

Most database administration officials concern themselves with engineering solid code, but there’s a physical side to the story as well. Database designers have to consider physical storage methods whenever they put together a new database. No one wants to see an overflow error, and this means that everyone needs to make good engineering decisions from the very beginning.

Physical Storage Systems

Database servers consist of a group of dedicated computers that hold the actual data that’s stored in the system. They usually only run the database management system package that the organization uses. Most servers will have nothing else installed on them except for their own operating systems and some related software.

Most modern database servers are built on multiprocessor architecture. They have generous amounts of RAM, though the actual amount necessary will depend on the type and size of database being used. Most engineering experts go out of their way to overestimate the amount of RAM necessary for access. Memory overflow errors are extremely disruptive to database users, which might help to explain these engineering decisions.

A majority of modern database servers rely on RAID data storage virtualization technology. Engineering textbooks usually define the acronym as redundant array of independent disks, but there seems to be some disagreement on the original meaning. The technology combines multiple disk drive components into a single logical unit.

This allows for data redundancy schemes to be put into play, which helps to ensure that no part of the database gets deleted. Computers writing to a RAID drive see it as a single disk. The array itself may make numerous copies of information written to the drive, but the host computer will see only discrete dataset.

A majority of RAID systems employ some sort of error protection scheme. This scheme, known as parity, is a widely used method to prevent errors within a certain level of fault tolerance. Some clever engineering students have developed systems that also provide additional security by combining physical hard disks with solid-state drives. The two technologies mirror each other in these arrays, which can help to take advantage of the benefits of both while minimizing the risks of either.

Logical Database Structure

Software engineering personnel will also have to make some decisions regarding how the database structure is coded. They can do this in a few different ways. Once the dependencies among each piece of information have been determined, it’s possible to arrange data into some sort of a logical structure. This structure can then be mapped into the storage objects that the database management systems provide.

Relational databases are the easiest for new developers to understand. These take storage objects and put them into tables that store data in rows and columns. Each table represents either the implementation of a logical object or some sort of relationship that joins instances of one or more logical objects together.

Tables can have relationships between each other as well. The can be stored as links that connect child tables with larger parent tables. Complex logical relationships are tables in their own right, and they can ultimately have links to more than one parent as well.

An object database features storage objects that directly correspond to objects used in programming code. An object-oriented programming language is a requirement in these sorts of schemes. The language is used to write the applications that will ultimately manage and access the data later on. The relationships can be defined as attributes of object classes that are involved. They could alternatively be defined as methods that operate on those object classes.

Big Data and Data Analytics

In recent years database administrators have faced the task of managing and storing big data. With the increase in data mining and development of the Internet of Things, very large data sets are being collected globally. Database administrators help pass information to data scientists in order to examine data for hidden patterns, market trends, customer preferences, and correlations. This process is named big data analytics and is a growing field with focuses in predictive analytics, data mining, text analytics, and statistical analysis.

Role of the Database Administrator

Database administrators are engineering professionals who take on the responsibility of almost every aspect of database management in an organization. This can involve systems monitoring, installing, configuration and maintenance. Many database administrators have to install updates and make sure that security patches get activated on time.

Many engineering experts will have to generate reports every time one is needed. They generally do this by querying information from the database. They also have to make sure that backups are completed on time. Reports seem to indicate that few people are actually familiar with how to restore a database from backups, and this means that engineering officials are also called on when something goes wrong.

When new users need to be enrolled in a database access scheme, they will have to go to the database administrator to request credentials. Database administrators usually maintain a list of all of the people who have existing credentials, which can help to cut down on unauthorized access to a database structure.

Application developers will sometimes request changes to the database structure to better fit the projects that they’re working on. They’ll call on the database administrator to modify the structure, and then the database will have to be tested to ensure that it’s still solid even after the changes are made. Altering a database can result in unintended consequences, which is why these tests are so necessary.

One of the biggest tasks that database administrators face is system storage allocation. Administrators have to allocate storage for the initial database array, and they have to plan for future storage requirements. Figuring out how large a database structure will possibly be months down the road is a very difficult task. Many databases grow at huge rates, which can help to explain why overestimation of requirements has become so common when making engineering decisions.


Sources

http://pages.cs.wisc.edu/~remzi/OSTEP/file-raid.pdf
http://www.drdobbs.com/database/sets-data-models-and-data-independence/228700616
LEARN MORE TODAY