UNIT 1: INTRODUCTION TO DISTRIBUTED DATABASES

Hrs: 8 | Marks: 8

Topics & Subtopics

●​ Distributed Data Processing​

○​ Centralized vs Distributed Systems​

○​ Client-Server, Peer-to-Peer, Multi-tier architecture​

○​ Goals of distributed data processing​

●​ What is a Distributed Database System (DDBS)?​

○​ Definition, Components of DDBS​

○​ Distributed DBMS vs Centralized DBMS​

●​ Advantages and Disadvantages of DDBS​

○​ Advantages: Reliability, Availability, Scalability, Autonomy​

○​ Disadvantages: Complexity, Overhead, Security, Data integrity​

●​ Problem Areas in DDBS​

○​ Data distribution​

○​ Query processing​

○​ Concurrency and Recovery​

○​ Security & Integrity​

●​ Overview of Database & Computer Network Concepts​

○​ Basics of Relational Databases (Normalization, ER to Relational)​

○​ Basics of Networking (LAN, WAN, TCP/IP, OSI model)​

UNIT 2: DDBMS ARCHITECTURE, DESIGN & QUERY

Distributed DBMS Architecture

●​ Types of Transparencies​

○​ Distribution, Transaction, Performance, Fault Tolerance, Replication​

●​ Architectures​

○​ Client-server​

○​ Peer-to-peer​

○​ Federated systems​

●​ Global Directory Issues​

○​ Centralized vs Distributed Directory​

○​ Directory management and consistency​

Distributed Database Design

●​ Alternative Design Strategies​

○​ Top-down, Bottom-up, Mixed approach​

●​ Fragmentation​

○​ Horizontal, Vertical, Mixed​

○​ Correctness criteria: Completeness, Reconstruction, Disjointness​

○​ Centralized, Replicated, Partitioned​

○​ Allocation strategies (Optimal, Heuristic)​

Semantics & Data Control

●​ View Management​

○​ Virtual views in distributed systems​

●​ Data Security​

○​ Access control, Encryption, Secure channels​

●​ Semantic Integrity Control​

○​ Constraints (PK, FK, Check), Global constraints, Distributed triggers​

Query Processing Issues

●​ Objectives​

○​ Minimize response time, Reduce communication cost​

●​ Characterization​

○​ Query processor components (parsers, optimizers, executors)​

●​ Layers of Query Processing​

●​ Query Decomposition​

○​ Query parsing, rewriting, and fragment identification​

●​ Localization of Distributed Data​

○​ Data localization using fragments​

Distributed Query Optimization

●​ Factors Governing Optimization​

○​ Communication cost, Processing cost, Local site constraints​

●​ Centralized Query Optimization​

○​ Cost-based, Rule-based approaches​

●​ Ordering of Fragment Queries​

○​ Semi-join strategy, Join ordering​

●​ Distributed Optimization Algorithms​

○​ Dynamic programming, Greedy heuristics​

○​ Algorithm to study: Selinger's Cost-Based Optimization​

Transaction Management

●​ Transaction Concept​

○​ ACID properties​

●​ Goals​

○​ Atomicity, Consistency, Isolation, Durability​

●​ Characteristics​

○​ Read/write sets, Conflicts, Schedules​

●​ Taxonomy of Transaction Models​

Concurrency Control

●​ In Centralized Systems​

○​ Lock-based protocols, Timestamp ordering, MVCC​

●​ In Distributed Systems​

○​ Distributed 2PL, Timestamp ordering​

●​ Distributed Algorithms​

○​ Two-Phase Locking (2PL)​

○​ Wound-Wait & Wait-Die​

●​ Deadlock Management​

○​ Prevention, Detection, Resolution​

○​ Global Wait-For Graph​

UNIT 4: RELIABILITY AND RECOVERY

Reliability Issues

●​ Reliability vs Availability​

Types of Failures

●​ Transaction failures​

●​ System (site) failures​

Reliability Techniques

●​ Redundancy, Checkpointing, Logging​

Commit Protocols

●​ Two-Phase Commit (2PC)​

●​ Three-Phase Commit (3PC)​

Recovery Protocols

●​ Distributed recovery​

●​ Log-based recovery​

●​ Cascading rollback, Checkpointing​

UNIT 5: PARALLEL DATABASE SYSTEMS

Parallel Architectures

●​ Shared Memory, Shared Disk, Shared Nothing​

●​ Interconnection networks (Bus, Mesh, Hypercube)​

Parallel Query Processing

●​ Intra-query and Inter-query parallelism​

●​ Pipelining vs Partitioning​
 ●​ Granularity of parallelism​

UNIT 6: ADVANCED TOPICS

Mobile Databases

●​ Characteristics: limited bandwidth, mobility​

●​ Data synchronization, Caching, Disconnection management​

Distributed Object Management

●​ Object-oriented DBMS concepts​

●​ CORBA, Distributed object references​

Multi-databases (Federated DBs)

●​ Local autonomy​

●​ Schema integration​

●​ Query processing in multi-DBMS​

🔑 Summary of Important Algorithms

Fragmentation Fragmentation correctness checks

Concurrency 2PL, Timestamp Ordering, Wait-Die, Wound-Wait

Deadlock Global Wait-For Graph Analysis

Commit 2PC, 3PC

Recovery Log-based Recovery (ARIES if covered),