How do new database models provide better scalability and performance over traditional RDBMS implementations?

Question

We have

• BigTable from Google ,
• Hadoop, actively contributed by Yahoo ,
• Dynamo from the Amazon

aimed at achieving one common goal - providing the greatest possible scalability of data management .

Due to scalability, I understand that the cost of use should not increase dramatically with increasing data size.

RDBMS are slow when the amount of data is large, as the number of unchanging immutable goals leads to more IO.

How do these customizable, scalable, user-friendly data management systems solve the problem?

, Google BigTable:

. ?

Answer 1

"" SQL , - . . , -.

, "" . , . , , , , () . , SQL , NOSQL ( SQL).

Answer 2

Bigtable, , - , . Bigtable tabletserver ( ); , Bigtable. ( , ), , , .

, , , , : , , , , ( , !).

Answer 3

, comodity // (, /).

RDBMS (), . , DB, ..

, - , .

- / , , (?) ​​ .

- - (), .

Answer 4

- http://queue.acm.org/detail.cfm?id=1394128 - ACID . . http://database.cs.brown.edu/papers/stonebraker-cacm2010.pdf -.

. MySQL Oracle. MySQL , Oracle, . MySQL DB, , , Oracle, , MySQL Berkely DB , , , ld, , .