Memory & Storage: Types, Hierarchy & Data Management

TAKEAWAYS:

• Memory hierarchy (CPU registers → cache → RAM → SSD → HDD → cloud) balances speed, cost, and capacity using a 6-layer pyramid (visualized below).
• Volatile vs. non-volatile storage: RAM loses data on power-off; HDD/SSD retain data permanently (critical for tourism booking systems like eSewa or Daraz).
• Data representation: Binary (0/1) and hexadecimal (0-F) encode text, numbers, and commands (e.g., a hotel room booking ID A1B2 = 10100001 00011010 10110010 00100010 in binary).
• Storage devices: HDDs use magnetic platters (slow, cheap); SSDs use NAND flash (fast, expensive); cloud storage (e.g., Google Drive) distributes data across servers.
• Fragmentation: Over time, files split into non-contiguous clusters, slowing access (fix with de-fragmentation tools like Windows Disk Defragmenter).
• Backup strategies: RAID (Redundant Array of Independent Disks) mirrors data for reliability (used by banks like Nabil Bank for transaction logs).

1. Memory Hierarchy: The Speed-Cost-Capacity Trade-off

Computers use a multi-layered memory system to balance speed, cost, and storage capacity. The closer the memory is to the CPU, the faster but more expensive it is. Here’s the hierarchy with real-world analogies:

Why this matters for tourism?

• Primary memory (RAM): Stores active data like a hotel reservation system’s live bookings (e.g., Yatra.com or Nepal Tourism Board’s website). If RAM fills up, the system slows down or crashes.
• Secondary storage (HDD/SSD): Holds historical booking data, customer profiles, and payment records (e.g., Khalti’s transaction logs).
• Cloud storage: Used by global platforms like Booking.com to store petabytes of data across servers worldwide.

Worked Example: eSewa Payment Gateway When you book a flight on Yatra.com and pay via eSewa, the process involves:

1. CPU registers/cache: Temporarily hold your card details during encryption.
2. RAM: Stores the transaction in progress (e.g., User_ID: 12345, Amount: NPR 5000, Status: Pending).
3. SSD: Saves the transaction permanently in eSewa’s database.
4. Cloud backup: A copy is stored in AWS or Google Cloud for disaster recovery.

2. Primary Memory: RAM and ROM

Primary memory is directly accessible by the CPU and is divided into two types:

A. RAM (Random Access Memory)

• Volatile: Loses data when power is off.
• Types:
  • DRAM (Dynamic RAM): Cheaper, slower (used in desktops).
  • SRAM (Static RAM): Faster, expensive (used in CPU cache).
• Capacity: Measured in GB (e.g., 8GB RAM).
• Speed: Accesses data in nanoseconds (ns).

Real Picture of RAM:

Why RAM is critical for tourism software?

• Example: A travel agency’s GDS (Global Distribution System) like Amadeus or Sabre uses RAM to process real-time flight/hotel searches. If RAM is insufficient, searches slow down, losing customers.

B. ROM (Read-Only Memory)

• Non-volatile: Retains data even when powered off.
• Types:
  • PROM: Programmable once.
  • EPROM/EEPROM: Erasable and reusable.
  • Flash Memory: Used in USB drives and SSDs.
• Use in tourism: Stores firmware for POS machines (e.g., Khalti’s payment terminals) or BIOS in airport check-in kiosks.

Comparison Table: RAM vs. ROM

3. Secondary Storage: HDD vs. SSD vs. Cloud

Secondary storage permanently holds data even when the computer is off. Key devices:

A. HDD (Hard Disk Drive)

• How it works:
  • Uses magnetic platters and a read/write head to store data.
  • Slower (50–150 MB/s) but cheaper (e.g., $0.03/GB).
• Used in:
  • Older servers (e.g., NTC’s legacy billing systems).
  • Backup drives (e.g., Nepal Rastra Bank’s audit logs).

Real Picture of HDD:

B. SSD (Solid State Drive)

• How it works:
  • Uses NAND flash memory (no moving parts).
  • Faster (300–3500 MB/s) but expensive (e.g., $0.10/GB).
• Used in:
  • Laptops for travel agents (e.g., MacBook Air).
  • High-speed databases (e.g., Nepal Stock Exchange’s trading system).

Worked Example: NEPSE Trading System

• HDD: Too slow for real-time stock price updates.
• SSD: Used to store live trading data and execute orders in milliseconds (critical for high-frequency trading).

C. Cloud Storage

• How it works:
  • Data is stored on remote servers (e.g., Google Cloud, AWS).
  • Accessed via internet (latency depends on connection speed).
• Used in:
  • Global travel platforms (e.g., Booking.com, Expedia).
  • Nepali companies (e.g., Daraz’s inventory management).

Comparison Table: HDD vs. SSD vs. Cloud

4. Data Representation: Binary and Hexadecimal

Computers store all data in binary (0s and 1s). For humans, hexadecimal (0-F) is used for compact representation.

A. Binary Numbers

• Example: The number 5 is 0101 in binary.
• Text storage: Uses ASCII or Unicode (e.g., A = 65 in decimal = 01000001 in binary).

Worked Example: Hotel Room Booking ID

• Suppose a hotel assigns room IDs as A1B2.
• Binary representation:
  • A = 10100001
  • 1 = 00000001
  • B = 10110010
  • 2 = 00000010
• Combined: 10100001 00000001 10110010 00000010

B. Hexadecimal Numbers

• Base-16 system (0–9, A–F).
• Example: A1B2 in hex = 10100001 00000001 10110010 00000010 in binary.

Why hexadecimal?

• Used in memory addresses (e.g., 0x7FFE).
• Networking: MAC addresses (e.g., 00:1A:2B:3C:4D:5E) are hexadecimal.

5. Storage Management: Fragmentation and RAID

A. Fragmentation

• Over time, files are split into non-contiguous clusters, slowing access.
• Example: A 1GB file might be stored in 100 different clusters across an HDD.
• Solution: Defragmentation (e.g., Windows Disk Defragmenter).

Mermaid Diagram: Fragmented vs. Defragmented File

B. RAID (Redundant Array of Independent Disks)

• Combines multiple disks for speed, redundancy, or both.
• Types:
  • RAID 0: Striping (faster, no redundancy).
  • RAID 1: Mirroring (redundancy, same capacity as one disk).
  • RAID 5: Striping + parity (balance of speed and redundancy).

Real-World Use in Tourism:

• Nabil Bank’s ATM network uses RAID 1 to ensure transaction logs are never lost.
• Airport security systems use RAID 5 for real-time passenger data.

6. Backup Strategies for Critical Data

Tourism businesses (hotels, airlines, travel agencies) cannot afford data loss. Common backup methods:

1. Full Backup: Copy all data (time-consuming but comprehensive).
2. Incremental Backup: Only new/changed data since last backup.
3. Differential Backup: All changes since the last full backup.

Example: Nepal Airlines Reservation System

• Primary storage: SSD for live bookings.
• Backup: Cloud storage (AWS) + RAID 1 for redundancy.
• Disaster recovery: Offsite backups in Pokhara data center.

In the Real World

1. eSewa and Khalti Payment Gateways

   • Idea Used: RAM for temporary transaction storage and SSD/cloud for permanent records.
   • How: When you pay for a flight on Yatra.com, the transaction is first stored in RAM (volatile) for processing, then moved to SSD/cloud (non-volatile) for permanent records. If the power goes out during processing, the transaction is lost (hence the need for backup power supplies in data centers).

2. Daraz’s Inventory Management

   • Idea Used: Database indexing on SSDs for fast product searches.
   • How: Daraz uses SSDs to store product catalogs (e.g., Product_ID: 12345, Name: "Nike Shoes", Price: NPR 5000). When you search for "running shoes," the SSD retrieves results in milliseconds instead of seconds (as HDDs would).

3. Nepal Stock Exchange (NEPSE) Trading System

   • Idea Used: RAID 0 for speed and RAID 1 for redundancy.
   • How: During trading hours, NEPSE’s system uses RAID 0 to process thousands of orders per second. At the end of the day, a RAID 1 mirror ensures all trades are backed up instantly in case of a crash.

Exam Tip

1. Memory Hierarchy: Always draw the 6-layer pyramid (CPU registers → cache → RAM → SSD → HDD → cloud) and explain the speed-cost-capacity trade-off. Examiners love this visual!

   • Example Question: "Why does a computer use both RAM and HDD?"
   • Answer: RAM is fast but volatile (for active tasks like live bookings), while HDD is slow but non-volatile (for permanent storage like customer databases).

2. RAM vs. ROM: Know the key differences (volatile/non-volatile, speed, cost) and give real-world examples (e.g., RAM for eSewa transactions, ROM for airport kiosk firmware).

3. HDD vs. SSD: Compare speed, cost, and durability with tourism examples (e.g., SSDs for NEPSE trading, HDDs for NTC billing backups).

4. Data Representation: Be able to convert decimal to binary/hex and vice versa. For exam questions, show step-by-step conversion (e.g., A1B2 → binary).

   • Example Question: "Convert the hexadecimal value 1A3F to binary."
   • Answer:

     1 = 0001
     A = 1010
     3 = 0011
     F = 1111
     Combined: 0001 1010 0011 1111
     

5. Fragmentation and RAID: Explain why fragmentation slows down storage and how RAID improves reliability. Use banking or airline examples (e.g., Nabil Bank’s RAID 1).

6. Backup Strategies: Know the difference between full, incremental, and differential backups and why tourism businesses need them (e.g., hotel reservation crashes during peak season).

Final Visual Summary:

graph LR
  A["CPU"] -->|"Fastest"| B["Cache"]
  B --> C["RAM"]
  C -->|"Volatile"| D["Loses data on power off"]
  C -->|"Slower"| E["SSD"]
  E -->|"Non-volatile"| F["Retains data"]
  E -->|"Slower"| G["HDD"]
  G -->|"Slowest"| H["Cloud"]
  H -->|"Redundant"| I["RAID/Backups"]