Mobile Computing: Devices, Networks & Challenges

What is Mobile Computing?

Mobile computing refers to the use of portable computing devices (smartphones, tablets, wearables) connected to wireless networks to access and process information anytime, anywhere. Unlike traditional computing, it emphasizes mobility, connectivity, and context-awareness.

Key Components of Mobile Computing

classDiagram
  class MobileDevice {
      +CPU
      +RAM
      +OS
      +Battery
      +Sensors
      +Display
      +Camera
  }
  class WirelessNetwork {
      +Wi-Fi
      +4G/5G
      +Bluetooth
      +NFC
      +Satellite
  }
  class MobileApp {
      +Native
      +Hybrid
      +Web
  }
  MobileDevice "1..*" --> "1" WirelessNetwork : "Connects via"
  MobileDevice "1" --> "1..*" MobileApp : "Runs on"
  WirelessNetwork "1" --> "1..*" MobileApp : "Enables communication"
  note for MobileDevice "Examples: Samsung Galaxy, iPhone"
  note for WirelessNetwork "Examples: Nepal Telecom, Ncell"
  note for MobileApp "Examples: WhatsApp (Hybrid), Instagram (Native)"

Types of Mobile Devices

Mobile devices vary by form factor, processing power, and use case. Below is a comparison:

Real-World Example:

• eSewa (Nepal) uses smartphones for mobile payments, leveraging NFC (Near Field Communication) for contactless transactions. The app connects to 4G/5G networks for real-time bank synchronization.
• Pathao (ride-hailing) relies on GPS sensors in smartphones to track driver locations and 5G networks for low-latency ride requests.

Mobile Device Architecture

Mobile devices follow a layered architecture, similar to traditional computing but optimized for power efficiency and portability.

Hardware Layers

graph TD
    A["SoC (System on Chip)"] --> B["CPU: ARM Cortex"]
    A --> C["GPU: Mali/Adreno"]
    A --> D["NPU: Neural Processing Unit"]
    A --> E["RAM: LPDDR"]
    A --> F["Storage: eMMC/UFS"]
    A --> G["Battery: Li-ion"]
    A --> H["Sensors: Accelerometer, Gyroscope, Proximity"]
    A --> I["Connectivity: Modem, Wi-Fi, Bluetooth"]

Software Layers

flowchart LR
  A["Hardware"] --> B["Baseband Processor"]
  B --> C["Android/iOS Kernel"]
  C --> D["Middleware: HAL, Android Runtime, Binder IPC"]
  D --> E["Framework: Activities, Services, Content Providers"]
  E --> F["Apps: Native/Hybrid"]
  E --> G["System Apps: Phone, Camera"]
  F -->|"API Calls"| E

Worked Example: How a WhatsApp Call Works

1. User initiates call → Microphone captures audio (hardware).
2. Audio processed by SoC’s DSP (Digital Signal Processor).
3. Encrypted via Signal Protocol (software layer).
4. Sent over 4G/5G (network layer) to recipient’s device.
5. Decrypted and played on recipient’s speaker.

Visual Trace of Data Flow:

sequenceDiagram
    participant User
    participant Mic
    participant SoC
    participant Network
    participant Recipient
    User->>Mic: Speak
    Mic->>SoC: Capture audio
    SoC->>SoC: DSP processing
    SoC->>Network: Encrypt & send (4G)
    Network-->>Recipient: Receive
    Recipient->>Recipient: Decrypt & play

Wireless Networks in Mobile Computing

Mobile devices rely on heterogeneous wireless networks for connectivity. Below is a comparison of key technologies:

Real-World Example:

• NEPSE (Nepal Stock Exchange) uses 5G-enabled mobile apps for real-time stock trading. Low latency ensures traders execute orders faster than competitors.
• Daraz’s "Same-Day Delivery" relies on 4G/5G networks to track courier locations via GPS and update customers instantly.

Challenges in Mobile Computing

Mobile computing faces unique challenges due to limited resources and dynamic environments.

1. Battery Life

• Problem: Short battery life limits usage.
• Solutions:
  • Low-power modes (e.g., Android’s "Battery Saver").
  • Hardware optimizations (e.g., ARM’s big.LITTLE CPUs).
  • Software optimizations (e.g., WhatsApp’s background sync limits).

Visual: Battery Drain Over Time

2. Security and Privacy

• Threats:
  • Malware (e.g., fake banking apps on third-party stores).
  • Data leaks (e.g., unencrypted SMS in old apps).
  • Jailbreaking/rooting (bypasses OS security).
• Solutions:
  • Biometric authentication (fingerprint, Face ID).
  • App sandboxing (Android/iOS isolate apps).
  • Encryption (TLS for data in transit, AES for storage).

Worked Example: Khalti’s Security

1. User logs in → Fingerprint scanned (hardware).
2. App checks if device is rooted/jailbroken (software).
3. Transaction encrypted with AES-256 before sending to bank.
4. OTP verified via SMS (with SMS encryption).

3. Connectivity Issues

• Problems:
  • Network drops (e.g., Kathmandu traffic jams causing 4G handover delays).
  • Offline access needed for rural areas (e.g., NTC’s smart meters).
• Solutions:
  • Offline-first apps (e.g., Google Docs’ offline mode).
  • Adaptive bitrate streaming (YouTube adjusts quality based on speed).
  • Mesh networks (e.g., LoRaWAN for IoT in remote villages).

Visual: Network Handover in Kathmandu Traffic

sequenceDiagram
  participant Phone as User's Phone
  participant TowerA as 4G Tower (Nepal Telecom)
  participant TowerB as 5G Tower (Ncell)
  participant CoreNetwork as Mobile Core Network
  Phone->>TowerA: 4G Signal (RRC Connected)
  loop Handover Process
      TowerA-->>Phone: Signal Strength: -85 dBm
      Phone->>CoreNetwork: Measurement Report
      CoreNetwork-->>TowerB: Handover Command
      TowerB-->>Phone: Sync & Re-establish
  end
  Phone->>TowerB: 5G Signal (NR Connected)
  note right of Phone
      Handover Latency: ~50ms
      Common in Kathmandu's Thamel
      to Putalisadak traffic areas
  end

4. Fragmentation

• Problem: Many device models (e.g., 100+ Android versions) lead to inconsistent app performance.
• Solutions:
  • Cross-platform frameworks (Flutter, React Native).
  • Feature detection (e.g., check if device has NFC before using it).
  • Beta testing on multiple devices (e.g., Pathao tests on Redmi, iPhone, and Samsung).

Mobile Computing vs. Traditional Computing

In the Real World

1. eSewa (Mobile Payments)

   • Idea Used: NFC + Secure Element (SE) chips in smartphones.
   • How: When you tap your phone on a POS machine, the SE chip (a secure hardware module) stores your payment credentials and processes the transaction without exposing data to the app. This prevents hackers from stealing data even if the phone is compromised.
   • Nepal Context: During Dashain, eSewa sees 10x traffic as people transfer money for gifts. The system must handle thousands of transactions per second using load-balanced servers and CDN caching.

2. Pathao (Ride-Hailing)

   • Idea Used: Real-time GPS tracking + 5G edge computing.
   • How: When you request a ride, Pathao’s app:
     • Uses GPS sensors to pinpoint your location (accuracy: ±3m).
     • Sends this data to 5G edge servers (located near towers) for low-latency processing (vs. cloud servers in the US).
     • Matches you with the nearest driver using geohashing (divides Kathmandu into grids).
   • Nepal Context: During Chhath festival, Pathao’s servers must handle 50% more requests due to temple crowds. They use auto-scaling (adding more servers dynamically) to avoid crashes.

3. NTC’s Smart Meters

   • Idea Used: LoRaWAN (Low-Power Wide-Area Network).
   • How: Traditional meters require a technician to read them monthly. NTC’s smart meters:
     • Use LoRa chips (consume 1/10th the power of Wi-Fi).
     • Send data every 2 hours via sub-GHz frequencies (penetrate walls better than 4G).
     • Connect to NTC’s central server via base stations (no need for 4G towers everywhere).
   • Nepal Context: In remote villages like Humla, where 4G towers don’t exist, LoRaWAN ensures 99% meter accuracy without manual checks.

Exam Tip

This unit is conceptual but heavily tested on applications. Expect:

1. Definitions: Be ready to explain terms like SoC, NFC, edge computing, and fragmentation in 1–2 sentences.
2. Comparisons: Questions will ask you to compare 4G vs. 5G, Android vs. iOS architectures, or Wi-Fi vs. Bluetooth.
3. Scenario-Based: You’ll be given a real-world situation (e.g., "Why does Pathao fail in rural areas?") and must link it to network coverage, battery life, or hardware limitations.
4. Diagrams: Always draw layered architectures (hardware/software) or data flow sequences (e.g., WhatsApp call) in exams. Label every component.
5. Short Answer: Memorize 3 key challenges (battery, security, connectivity) and 2 solutions for each.

Common Pitfalls:

• Confusing 4G latency (50ms) with 5G latency (1ms). Always state exact values.
• Forgetting Nepal-specific examples (e.g., NTC’s LoRaWAN, eSewa’s NFC). Examiners love local cases.
• Describing Wi-Fi as a mobile network—it’s not (mobile implies cellular data).

Pro Tip: For the exam, practice explaining how a single app (e.g., Khalti) uses 3 different mobile computing concepts (e.g., NFC for payments, GPS for location, and 4G for syncing). This shows depth of understanding.