History

Flat File System

A Flat File is a database that stores the data in a plain text file. Each line of the file stores a single value. Each field is separated by delimiters such as commas or tabs. Although it can have multiple tables, it cannot have multiple relations as the Relational Databases have.

Relational Databases

The main advantage of Relational Database over the Flat File System is that the proper organization of data that has. A relational database also defines the relationships between those tables. In the relational databases, queries are used to fetch data with the help of the indexes. The relational database technology makes the databases efficient, lighter and faster. Some of the advantages of the relational databases are;

• Bringing tables together using relations
• Provides a structure query language (SQL) to define and manipulate data.
• Security

Object Relational Model: It is also a relational data model but with object orientation in it. It reduces the gap between the conceptual data modeling techniques and object-relational mapping.

Object-Oriented Database

An Object Oriented database is a combination of objects in a persistent storage which holds information. It is quite similar to the object oriented languages. It can be named as the fifth-generation database technology that was began to develop in mid 80’s.  The real world entities are represented like an object in the Object Oriented Data Model.

Multidimensional Database

It is a database system that can be used to utilize the advantages of the databases. It is usually structured to optimize online analytical processing and data warehouse applications. The multidimensional database can receive data from a variety of relational databases and structure the information into categories and sections that can be accessed in number of different ways.

Sub Modules

Engine

There are several tasks done by the DBMS. It accepts various logical requests from various other DBMS sub systems. Then convert those requests into their physical equivalent and actually access the database and data dictionary as they exist on a storage device.

Data definition and manipulation

In Structured query languages, they have four different operations mainly to operate data. They are Select (Retrieve data from a table/tables), insert (insert data to the tables), update (update a certain field of a column) and delete (delete rows). In addition to them there are several other commands for creating, dropping databases, etc…

Administration

The database administration means the maintaining the functionalities and managing the DBMS. There are 3 main types of Administration. They are

• Operation Administration
• Development Administration
• Application Administration

There comes several tools for the database administration and they are called as native tools.

Other services

Security: The security of data is the most important thing. A DBMS allows to let several users to access the database up to several extends. In addition to that the DBMS allows setting passwords. In addition to that the DBMS allows setting passwords. Encrypting the data is another security feature provided by the advanced DBMS.

Disaster recovery: Database Management Systems allows its users to maintain backups. It makes easy to recover when there is a crash of data. In addition to that most of the DBMS provide role back to a previous state in order to undo transactions in a given period..

Indexing: Indexing makes the operations such as insert and update easy. The data is ordered randomly and the way how it was ordered is described by the logic Indexing speed up manipulation of data since it keep a summary of mata date and it make the SQL engine to access the required data without searching whole database.However indexing doesn’t seep duo the data manipulation in the case the query is conditional, like ‘select name from student where age<10’.  In such a scenario SQL engine must go through whole table regardless of the indexed data.

Advantages of Using A DBMS

• Application programs are independent of data representation and data storage details.
• Data is stored and accessed efficiently including the support of very large files and index structures.
• Centralize administration of data shared among many users.
• Let users feel the data is accessed only by him at that time.

Selecting DBMS

Selecting an appropriate DBMS is an important decision in any software application. An inappropriate application in DBMS will result system instability and lower performance. In the software development context is similar to below cases it’s recommended to not to use a DBMS.

• Single-user applications PC applications.
• Non mission critical systems.
• Non security critical systems.
• If the application handles low volume of data
• In low resource environments. (Ex: Mobile Devices)

Current Database Management Systems

There are several Database Management Systems have been introduced as it became much popular. Some of the most popular databases are MySQL, Microsoft SQL Server, Oracle, Dataphore etc, and they have their own unique features. Considering them software developers can use them for their projects.

Oracle

• Oracle has many advanced functions such as Real Cluster, Flush back recovery etc.
• Oracle DBA and developer cost higher to hire and train. Oracle price is higher and charge you 22% support/maintains every year.
• For very complicated store procedure, Oracle runs faster.
• Oracle runs on many platforms. SQL Server only runs on Windows.
• If you have a big database needs running on powerful Unix or IBM server, with complicated application. Oracle is a good choice.

MySQL

• Mysql replication works well for backup services and scaling out purposes are available.
• In Mysql master-slave replication model, changes only flow from master to slaves.

MSSQL

• T-SQL (Transaction SQL) enhancements
• CLR (Common Language Runtime)

Future Trends

Although the Database Management Systems have been very much developed, there are some areas yet to be developed. The areas like integrating structured and unstructured data, virtual access of data, and simplifying data management through greater automation and intelligence can be developed more than now they are. A research site estimates that more than 90% of all business data in enterprises is unstructured, but only 5% of unstructured data is stored in databases. The DBMS will address this challenge by supporting all kinds of data and content in its native form with much tighter integration, while sustaining high performance in the future as the experts expect.  In the next few years, DBMS technology and middleware will also evolve to support information fabric, virtual access to heterogeneous data.In addition to those, there will be a growth in e-Commerce and enterprise portals. Demand for commodity front-end database servers will increase and the market for the open source DBMS will be much higher than now.Mobile database suites will be a very good trend in the future and it will enable to access the databases located in the mobile phones.

References

• IBIBO masters. (2007, November 07). What is flatfile system. Retrieved May 07, 2010, from IBIBO: http://sawaal.ibibo.com/computers-and-technology/what-flat-file-system-198192.html
• Fundamentals of Database. In e. &. navathe, Fundamentals of Database,4th Eddition. Pearson Education, Inc.
• Yuhanna, N. (2005, September 29). The Future Of DBMS Technology. Retrieved May 07, 2010, from www.forrester.com: http://www.forrester.com/rb/Research/future_of_dbms_technology/q/id/37181/t/2
• insidetech.monster.com. (2009). Retrieved May 2010, from Get a Database Administrator Job: http://insidetech.monster.com/nfs/insidetech/attachment_images/0002/0929/database-250-2_crop380w.jpg

The problem in the previous system were

• Direct changeover
• Consequences in system failure
• No Driving fun

Let see if we will decentralize controlling, how it will solve all these problems. In this concept all the vehicles will be intelligent agents. Like every other human who are driving a vehicle sitting behind the wheel. In addition to that the cars can communicate with each other. In previous system all the decisions would be taken by the central controlling point. In here they will be taken by the cars themselves.

Ex: Car B is following car A. Car A going to turn left after 1 KM. Previous centralized system, that information will be uploaded to central server and it will make the car B aware. In this decentralized system car A will directly talk to car B and say “Hey dude! I’m going to turn left after 1KM” so based on that information car B can take steps.

Sounds cool ah! But still there can be a central server to update cars about general conditions of the roads including traffic and maintenance. And just assume, only your car is having this system. No problem, it can use its camera, radar etc and drive alone thinking all the other cars are driven by humans. With the increasing number of this kind of intelligent cars, system will come to live step by step. No need of direct change over. In the case of central server failure it won’t affect the system badly. Still cars can drive communicating each others.

That’s how this decentralized system is going to solve the problem of changeover and how it will react to a system failure. But how this is going to solve the problem of “Driving fun”?

There is a simple solution. “Manual override!”. If driver want to drive by himself, he would be able to override the system manually and continue driving. In that case all the other cars around will be notified that this car is manually overridden. So other cars can take appropriate security steps, because a human is behind the wheel. To make this even better, there can be some busy city areas where manual overriding is prohibited. In permitted areas drivers can have fun.

This sounds like an awesome system, but when it comes to implementation, there are many technical challenges. Still there is no good enough intelligent system to drive a car on a real road. Communication between cars wouldn’t be much harder. It’s just matter of introducing common protocol. Once we come up with a good intelligent system to drive a car, we are almost done.

(Special Thank : Shyam Mehraaj Reyal)

References

• (n.d.). Retrieved from http://www.mjnet.us/recen_t_projects_files/traffic_control_center_long_island_city_queens__traffic_moni.jpg

Just think, if everyone on the road will perfectly follow the road rules, In that case there won’t be any traffic. But humans are controlling vehicles and we are not that perfect. So is there any perfect one to control traffic. Yes, a computer. I actually saw this in the movie “Minority report”. This is a Centralized Traffic Control System. All the vehicles are connected to a centralized server and it control everything on the road. And the best part is the server knows positions of each and every vehicle, so it can move two vehicles nearby on the same track with the speed 120 Kmph, maintaining very narrow distance between them. And the same thing can be applicable to ‘n’ number of vehicles. On top of all, just assume the vehicle in front of you going to turn left in next junction. In general situation the driver on signal light in left and indicate that. But in this kind of system, the central server knows that the vehicle going to turn left and it can control other vehicles behind that accordingly. This sounds like a dream, but not impossible. We already have GPS and maps on our vehicles. All we have to do is update those details and get driving instruction from the server and execute them.

Once this kind of system will come to action, there won’t be a Steering wheel in vehicles. That space also will be allocated for passengers. The passenger will have to enter the destination once he gets in to the vehicle and those details will go to the central server. The server will give the vehicle a rout considering all the factors including other vehicle routs etc. Finally vehicle will reach the destination without human involvement. And it’s so difficult to find a parking space in modern busy cities and it’s a real pain to many people. But one day with this kind of system, once you go to office, you can tell your car to go home and come back when you going to finish the shift.

This will be a bad news for auto lovers. So sad that manual driving is no longer permitted. And you can’t drive over speed, break the law and have some fun. No more manual shifting, drifting, drag, street racing or anything. This is going to be kind a boring yes.

One possible problem in this kind of system is servers are always going down. Under any conditions we can’t guaranty 100% up time. What will happen if the central server (Probably the Server Cluster) will go down? Yes we can design the system to stop all the vehicles, but can we afford that delay?  On top of all, what will happen in the case of bug or defect in the system? The possible damage will be massive and probably cost many human lives. And how far can we expand this system? When it comes to implementation we have to go for a direct changer over. Is it possible? The engineers will have to find solutions for all those problems before this kind of system comes to live.

References

• (n.d.). Retrieved from http://www.wired.com/images/article/full/2007/10/FAA_630x.jpg