Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming.threads > #1963

Databases engines...

Cross-posted to 2 groups.

Show all headers | View raw

Hello,

I was thinking more about databases, and i have thought
to implement a database engine that supports INSERTS , DELETES and 
SEARCHES  and that supports indexed using the AND  and OR operator,
and that supports replication accross computers, not the
the master to slave replication but an enterprise master to master 
replication, but what will change in my database engine is
that the indexes of the database will stay in-memory, this will
speed the things more... but this will use more memory but this kind
of implementation , in term of how much memory it will take, will be 
between the the MySQL engine that uses much more the harddisk and an 
in-memory database... in my databse engine all the tables and indexes 
will be stored in the same hardisk file and i will use the same method 
as is using my ParallelVarFiler , that means that all the in-memory 
indexes will be saved automaticly to the hardisk...and the enterprise 
master to master replication will use TCP/IP and the database will be 
distributed and will use a client server TCP/IP mechanism and a 
distributed TCP/IP mechanism. I was saying to myself that my database 
engine's ideas , that i am presenting to you, are easy to implement, but 
i was thinking more about my database engine, but what i don't like in 
database engines , even the one that i want to implement, is that they 
are still using the a lot the hardisk and they are still disk bound, so 
if there is contention with many many users accessing the database this 
will degrade the throughput cause the database is still using a lot the 
hardisk and is disk bound...but to lower this impact of contention you 
have to distribute more your database accross many computers but this 
will be costly i think, other than that the realtime replication is 
costly also, so if we want to ensure a realtime replication and there is 
contention this will also degrade
the throughput i think. So since the databsae engines(even the one that 
i want to implement) have those weaknesses , so i have decided finally 
to not implement my database.


Hope you have understood my ideas.



Thank you,
Amine Moulay Ramdane.

Back to comp.programming.threads | Previous | Next | Find similar

Thread

Databases engines... aminer <aminer@toto.net> - 2013-10-25 22:42 -0700

csiph-web