A distributed database system consists of multiple servers, each responsible for its own data. These data can be accessed using a network and, despite being physically distributed, they must be logically integrated to the user, as if they were on a single server.
Why use Distributed Databases?
Need to maintain an integrated perspective of the organization's data.In other words, the data, despite being «physically distributed», must be «logically integrated».
(This is the main justification for distributed databases).
Properties of distributed databases.
In order to ensure data security and reliability, BDD preserves the ACID properties, where
A – Atomicity: Indicates that the transaction must have all its actions completed or not. If all actions in the transaction are successfully completed, then the entire transaction is Committed; otherwise the entire transaction must be rolled back, Rollback.
C – Consistency: Executing an isolated transaction preserves the consistency of the database.
I – Isolation: Each transaction is not aware of other concurrent transactions.
D – Durability: After a transaction is completed successfully (Commit), the changes it makes to the database persist.
Types of distributed databases.
Homogeneous Distributed Database.
Equal schemes
Similar hardware
Similar database management system
Heterogeneous Distributed Database
Different schemes
Different hardware
Different database management system
Different hardware and different database management system
Database distribution mechanisms
partition & replicate.
1.Partition.
By partitioning, the database is split, separated into several disjoint partitions.
2.Replicate.
In database replication, systems store copies of data at different sites.
Full replication and Partial replication
Synchronous and asynchronous replication
1.Synchronous replication
The disadvantage of this type of replication is that if a network failure occurs between the nodes that have replicas, neither the local nor the remote copies can be updated until the network failure is resolved.
Also maintaining multiple copies (replicas) of this data requires more system resources and makes managing it even more complex.
2.Asynchronous replication
The disadvantage of this form of replication is that you cannot fully mirror the master table.
In case it is updated and before the (periodic) update of the replicas is carried out, if you make a query to these same replicas, you will see the old data.
Advantages of using a distributed database.
- Information sharing.
- Faster access to data.
- Acceleration of query processing.
- Greater local autonomy.
- Greater availability.
- Greater reliability.
- Modular extensibility.
Disadvantages of using distributed database
- Greater complexity.
- Increased maintenance and communication cost.
- Security.
- Lack of standardization.
- Higher storage requirements.
- Maintaining integrity is very difficult.
- More complex database design.
Distributed database application areas.
- Airlines;
- Store Networks;
- Hotel Chains;
- Among others.
Tags
Databases