Janata Bank Ltd — SO (Assistant Network Engineer) — Written Exam: 2020প্রশ্ন ও সমাধান

An index is an auxiliary data structure (commonly a B+ tree) that maps key values to the physical location of rows, letting the engine skip full table scans and jump straight to matching rows.

How it works: an index stores (key, pointer) pairs sorted by key. A lookup does a fast tree search — O(log n) — instead of scanning every row — O(n).

Types:

• Primary (clustering) index — built on the primary key; data rows are physically ordered by this key (one per table).
• Secondary (non-clustering) index — built on non-key columns; points to row locations without reordering data (many allowed per table).
• Unique / composite / full-text indexes — specialized variants for uniqueness, multi-column lookups, and text search.

graph TD
  Q[Query: WHERE emp_id = 105] --> IDX[B+ Tree Index on emp_id]
  IDX --> P1[Pointer to Block 42]
  P1 --> ROW[Row: emp_id=105, Name=Rafi]

Uses of indexing:

• Drastically speeds up SELECT/WHERE/JOIN/ORDER BY by avoiding full scans.
• Enforces uniqueness cheaply (unique index).
• Supports fast sorting, since index entries are already ordered.
• Reduces I/O for range queries (BETWEEN, <, >).

Trade-off: every index adds storage and slows down INSERT/UPDATE/DELETE (index maintenance), so indexes are added selectively on frequently-queried columns.

Exam takeaway: Index = B+ tree lookup structure trading write cost and storage for much faster reads; a primary index reorders data, a secondary index does not.

Data loss prevention combines backup discipline, redundancy, and security controls so no single failure — hardware crash, ransomware, human error, or theft — destroys the only copy of important data.

Key measures:

• Regular, scheduled backups — automate full plus incremental backups; periodically test restores, not just the backup job.
• 3-2-1 backup rule — keep 3 copies of data, on 2 different media types, with 1 copy off-site/off-network (cloud or another physical location).
• Encryption — encrypt data at rest (including backup media) and in transit, so stolen media is useless without keys.
• Access control & device management — least-privilege accounts, remote-wipe for lost mobile devices, MDM policies.
• Anti-malware & email security — block phishing/ransomware before it can encrypt or exfiltrate data.
• RAID / replication — mirror disks (RAID 1) or replicate databases across nodes so a single-disk failure does not lose data.
• UPS / power protection — prevents data loss from sudden power failure during writes.
• Versioning — keeps historical versions so accidental overwrite/deletion can be rolled back.

graph LR
  A[Live Data] -->|daily| B[Local Backup]
  A -->|weekly| C[Off-site / Cloud Backup]
  B --> D{Restore Test}
  C --> D

Exam takeaway: No single safeguard is enough — layer backups (3-2-1 rule), encryption, access control, and redundancy (RAID/replication) so any one failure remains recoverable.