JNTU MCA 2021 DISTRIBUTED DATABASES Previous Question Paper (July/August - 2021 - R17)

A Distributed Database is a collection of multiple interconnected databases that are spread physically across various locations, such as different computers, servers, or network nodes, but are managed by a distributed database management system (DDBMS). From the user's perspective, the system appears as a single, unified logical database, masking the underlying physical distribution.

Compared to a traditional centralized database approach, a distributed database offers several distinct advantages:

• Reliability and Availability: In a centralized system, a single point of failure can render the entire database inaccessible. In a distributed system, if one site fails, the remaining sites can continue to operate, ensuring higher system availability and fault tolerance.
• Improved Performance: Data can be placed geographically closer to the users who access it most frequently. This reduces network latency and communication costs compared to accessing a single central server from remote locations.
• Modularity and Scalability: Distributed systems are easier to expand. New sites or nodes can be added to the network without requiring a complete overhaul of the existing infrastructure, allowing the system to grow incrementally as organizational needs increase.
• Local Autonomy: Each site in a distributed database can maintain a degree of local control over its own data, allowing for local administration, security management, and optimization, while still participating in the global database environment.
• Reduced Communication Costs: By processing data locally at the site where it is stored, the volume of data transmitted over the network is significantly minimized, leading to more efficient bandwidth utilization.

Transparency in a distributed database system is the separation of the higher-level semantics of the system from the lower-level implementation details. It allows users and application programs to interact with the database as if it were a single, centralized system, hiding the complexities of data distribution, fragmentation, replication, and hardware heterogeneity.

The layers of transparency are categorized as follows:

1. Data Distribution Transparency: This ensures that the user does not need to know where the data is physically located. It includes:
   • Fragmentation Transparency: Hides the fact that a relation is divided into fragments. Users query the logical relation, and the system maps it to the appropriate fragments.
   • Replication Transparency: Hides the existence of multiple copies of data. The system manages updates to all replicas automatically, ensuring consistency without user intervention.
   • Location Transparency: Ensures that the command used to access data does not depend on the location of the data or the site where the user is located.
2. Transaction Transparency: Ensures that transactions maintain ACID properties (Atomicity, Consistency, Isolation, Durability) in a distributed environment. This includes:
   • Concurrency Transparency: Allows multiple transactions to run concurrently without interfering with each other, typically managed through locking or timestamping protocols.
   • Failure Transparency: Ensures that the system remains operational and consistent even if individual nodes or communication links fail, often using protocols like Two-Phase Commit (2PC).
3. Performance Transparency: Ensures that the system performs efficiently regardless of the distribution. The query processor must optimize queries by choosing the best execution strategy, such as minimizing data transfer across the network.
4. Heterogeneity Transparency: Hides the differences in hardware, operating systems, and database management systems (DBMS) across various nodes, allowing a unified interface to access diverse data sources.