Hi all, I would like to know which one is fast. I thought that DBMS is fast compared to file system until my friend said that file system is fast. He also said that google, amazon does not use a DBMS to search something. It all started when I asked him about what are spider/crawler programs. But I dint ask him much on this one.
Can anyone tell me how searching a file system is fast that a DBMS?
Databases and file systems serve different purposes, so comparing them in terms of "speed" generally does not make much sense. For example, a fighter jet is undoubtedly faster than a bike, yet many people use bikes to get to work and hardly any uses fighter jets for the same purpose. (Yes, this example is a bit stretched.)
Databases provide much more functionality than file systems and, because of that, can sometimes appear slower, especially if you try to force a large gobs of unstructured data into them. However, file systems can be still significantly slower than databases. Most filesystems will be terribly slow if you try to store and manipulate thousands of files in a single directory, for example.
And while Google and Amazon do not use traditional RDBMS (the R - meaning relational - is especially important here), they are using other kinds of databases, sometimes called NoSQL databases. So no fast filesystem behind Google's indexes, actually. Again, these databases provide different functions and features from the traditional RDBMS and are not useful for everyone. A bank, for example, won't be keeping client records and transactions in a NoSQL database.
Another advantage of a database is that you only keep one copy of the data in a database. If you have duplicate files in a file system, they can be altered so you have two copies of the same data . . . which are different from each other