Bangladesh Oil Gas Mineral Corporation (PetroBangla) — Assistant Manager (CSE/IT) — Exam: 31.06.2024 (BUET)প্রশ্ন ও সমাধান

How to recover from a drive failure depends entirely on the RAID level in use:

RAID Level	Drive-failure recovery
RAID 0	No fault tolerance — data striped with no redundancy; a failed drive means data loss, recoverable only from backup.
RAID 1	Replace the failed drive; rebuild by copying data from the surviving mirrored drive.
RAID 5	Replace the failed drive; rebuild by recomputing lost data from the distributed parity + remaining drives.
RAID 6	Tolerates two simultaneous drive failures; replace the failed drive(s) and rebuild from double parity.
RAID 10 (1+0)	Replace the failed drive; rebuild from its mirrored pair (tolerates multiple failures as long as no mirrored pair loses both sides).

flowchart TD
    F[Drive fails] --> D{RAID level?}
    D -- RAID 0 --> L[Data lost - restore from backup]
    D -- RAID 1/10 --> M[Replace drive, rebuild from mirror]
    D -- RAID 5 --> P[Replace drive, rebuild from parity + remaining disks]
    D -- RAID 6 --> P2[Replace up to 2 drives, rebuild from double parity]

• General recovery steps: (1) identify and physically replace the failed disk, (2) the RAID controller automatically rebuilds the array using redundancy (mirror copy or parity computation), (3) monitor for a second failure during rebuild, which is the highest-risk window (especially for RAID 5).

Exam takeaway: RAID 0 has zero fault tolerance; RAID 1/5/10 tolerate one drive failure; only RAID 6 (and RAID 10 under favourable conditions) tolerates two.

An edge server is a server placed physically close to end users/devices, rather than in a distant centralised cloud data centre — the core enabler of edge computing, which processes data near its source.

graph LR
    D1[IoT Sensor] --> E[Edge Server - local]
    D2[Smart Camera] --> E
    E -->|only summarised/critical data| C[(Central Cloud)]

Need	Why it matters
Low latency	Local processing avoids the round-trip to a distant cloud — vital for autonomous vehicles, industrial control, gaming
Bandwidth savings	Only pre-processed/summarised data is sent onward, instead of raw high-volume streams
Reliability	Keeps operating during a cloud/internet outage, improving availability
Data privacy/security	Sensitive data (health, finance) can be processed and stay local
Scalability	Supports huge numbers of IoT devices without overwhelming the central cloud
Real-time analytics	Enables immediate decisions — e.g. predictive maintenance, smart traffic control

• Edge computing complements, rather than replaces, cloud computing: the edge handles time-critical/local processing, while the cloud still handles heavy long-term storage/analytics.

Exam takeaway: edge servers bring computation physically closer to the data source, cutting latency and bandwidth use while improving reliability and privacy — essential for real-time IoT applications.