Geek Answers Handbook

Are there easy-to-use distributed relational databases?

I talked to one of my friends about distributed relational databases using defintion:

  • Supports all relational databases (SQL, ACID)
  • Distributed (multiple servers)
  • Automatic (or automatic-ish) - Install the program on several servers, tell them any information they need for communication (IP addresses), and then figure out how to automatically distribute things.

(3) is a part that seems to be never performed anywhere. You can do sharding, but then you need additional code in your application to figure out which server to talk to. What I'm looking for is similar to Memcached, but as a relational database (Memcached is the key) and ACID-compatible (Memcached is only in memory and I don't have anything like transactions).

Obviously, something like this would be incredibly difficult to succeed, but it surprises me that I don't see to find any examples of this.

+7
source share
3 answers

Teradata provides this, but is sold as a device. They automatically have ACID shards and are compatible with SQL RDBMS, but you have to buy them from them.

+3
source

Caution: I represent NimbusDB.

NuoDB (formerly NimbusDB) does just that. This is difficult to do if you start with traditional database architectures, but get it for free if you start with modern distributed architecture.

+2
source

I think the problem is here ... What kind of replication do you need. There are many types of solutions:

http://wiki.postgresql.org/wiki/Replication,_Clustering,_and_Connection_Pooling

Part of replication is built into postgres 9.0 and above, btw.

+1
source

More articles:


All Articles