Security in Mobile Computing – Device Security
Evolution & Usage of Modern Mobile Devices
- Rapid hardware & software evolution → smartphones = powerful general-purpose computers in pockets.
- User behaviour shift:
- >70\% of daily on-line time now on mobile (banking, shopping, e-mail, social media, IoT control, etc.).
- Device is simultaneously camera, wallet, key, personal diary, navigator.
- Security implications:
- Sensitive data is created, processed & stored locally.
- Always-on connectivity (Wi-Fi, 4G/5G, NFC, Bluetooth) broadens attack surface.
- Small size ⇒ easy to lose or steal ⇒ physical compromise risk.
- Monetisation channel built-in (operator billing, in-app purchases, NFC pay) ⇒ strong economic incentive for attackers.
Mobile Device Security Concerns
- Wireless communication is inherently broadcast ⇒ eavesdropping & spoofing easier than on wired links.
- Aggregated personal content on device:
- E-mails, calendars, contacts, pictures, location history, health data.
- Attack = single point of failure for user’s digital life.
- Growth of underground economy:
- Stolen data, botnet rental, ransomware, ad-fraud.
Personally Identifiable Information (PII)
- Definition: Any datum that can single-handedly or collectively isolate or identify a person.
- Typical mobile PII:
- E-mail address, phone number
- Hardware IDs: MAC, IMEI, IMSI, Bluetooth address
- Sensitivity & misuse:
- Tracking of movement, profiling of habits, linking of on-line & off-line identities.
- Machine-learning pipelines depend on PII for personalised recommendation; marketers maintain ever-growing databases.
- Security issues:
- \textbf{Insecure~transfer} – clear-text HTTP, weak TLS ⇒ interception.
- \textbf{Insecure~format} – PII stored as plain SQL/JSON/XML inside app sandbox.
- \textbf{Insecure~remote~storage} – poorly configured cloud DB ⇒ mass leak.
- \textbf{Accidental~exposure} – debug logs, query strings, dashboard widgets show e-mail / phone.
- Defensive mindset:
- Scrutinise permission requests; “least privilege” principle.
- Install from reputable stores only; deny excessive permissions.
Mobile Malware
- Definition: Malicious code aimed at compromising mobile OS, data or user wealth.
- Motivations & impacts:
- Harvest credentials ⇒ direct bank theft.
- SMS premium fraud – device silently texts pricey numbers.
- Botnet enlistment ⇒ DDoS, click-fraud.
- Data wipe / “bricking”.
- Infection vectors:
- \textbf{Phishing} – fake sites emulate originals.
- \textbf{Trojanised apps} – repackaged popular APK with extra payload; uploaded even to official stores.
- \textbf{Spyware} – background keylogging & screenshot.
- \textbf{Man-in-the-Middle (MitM)} on public Wi-Fi.
- \textbf{Direct attack} via MMS/SMS, Bluetooth worms.
- \textbf{Third-party repositories} – no curation.
- Example (Kakao Talk):
- Legit vs modified binary permissions list shows added abilities (send SMS, record audio, manage accounts) → red flag.
Best Practices to Avoid Malware
- Install only from official stores (Google Play, Apple App Store, Microsoft Store).
- Verify publisher identity; look for spoofed names & review counts.
- Avoid side-loading from *.apk, *.exe, uncertified markets.
- Review requested permissions; deny what is not required for core functionality.
- Keep OS & apps patched; enable Google Play Protect / iOS Gatekeeper.
Wireless Connection Security
Risks Associated with Text Messaging (SMS/MMS)
- Impersonation / identity theft – sender ID spoofing.
- Spam & malware delivery.
- Persistent local history ⇒ privacy loss upon device theft.
- Linguistic influence on youth (slang overuse) – social / educational concern.
Wi-Fi (WLAN) Security
- Wireless = broadcast medium: any radio within range can capture frames.
- Misconfigurations (open networks, weak WEP, default WPA keys) invite interception.
- Recommended counter-measures:
- Prefer encrypted networks; avoid open SSID.
- Use WPA2 or WPA3 with strong passphrase >!12 random characters.
- Employ VPN tunnelling (IPsec, WireGuard, OpenVPN) on public hotspots.
- Access only HTTPS sites (TLS \ge 1.2); watch for certificate warnings.
- Disable auto-connect to unknown SSIDs.
- Ethical note: Operators of free Wi-Fi often harvest traffic for analytics – informed consent & transparent policies needed.
“Evil Twin” Wi-Fi Attack
- Attacker sets up rogue AP named like legitimate (“UiTM_Hotspot”).
- Victim auto-connects; all traffic passes through attacker’s proxy.
- Enables credential stealing via fake captive portals or SSL-stripping.
- Defence: validate certificate, use VPN, inspect MAC/BSSID of AP, deploy 802.1X.
Bluetooth Security Threats
- Bluetooth forms Personal Area Network (PAN) automatically – convenience vs attack surface.
- Main attacks:
- \textbf{Eavesdropping / MitM} – attacker inserts two dummy devices between Victim A & Victim B, relays & records.
- \textbf{MAC spoofing} – clone victim’s MAC so inbound traffic diverts to attacker.
- \textbf{DoS – “Big NAK”} – flood with negative acknowledgements, draining battery.
- \textbf{Blue-Snarfing} – unauthorised reading/alteration of contacts, calendar, media.
- Best practice: keep Bluetooth off when not in use, pair only in private, use Bluetooth\,5 secure connections, set device to non-discoverable.
Authentication & Authorization
- Two-stage access control:
- \textbf{Authentication}: prove identity.
- \textbf{Authorization}: determine permitted actions.
Passwords / PINs
- Pros: cheap, ubiquitous; compatible with legacy systems.
- Cons: weakest factor; users choose predictable strings ("1234", "password").
- Administrative controls:
- Complexity policy (min length \ge 8, mix of classes).
- Expiry (e.g., 90 days).
- Account lockout after N failed attempts (usually N!=!5).
- Attack resilience: dictionary & brute force; mitigated by \text{salted hash} storage and key-stretching (PBKDF2, bcrypt).
Biometrics
- Body becomes the key. Traits:
- \textbf{Physiological} – fingerprint, face, iris, retina.
- \textbf{Behavioural} – voice, gait, keystroke dynamics.
- Advantages:
- Cannot be forgotten, difficult to share, harder to forge, requires presence.
- Challenges & ethics:
- Liveness detection (anti-spoofing), privacy of biometric templates, irreversible leak (fingerprints cannot be "revoked").
Pattern Lock (Android)
- 3 × 3 grid; user draws a swipe path length 4\text{–}9.
- Combinatorics: 389\,112 possible distinct patterns vs 456\,976 for 4-character case-insensitive password.
- Derivation (simplified): \sum_{k=4}^{9} P(9,k) where P(9,k)=\frac{9!}{(9-k)!}.
- Usability high; vulnerability: shoulder-surfing & oily residue.
- Hard & “cool” patterns exist, yet many users pick predictable shapes (letters, simple lines).
Comparative Summary
| Factor | Security | Ease of Use | Implementation | Works on Phones |
|---|---|---|---|---|
| Password/PIN | Weak | Easy | Easy | Yes |
| Biometrics | Strong | Moderate | Hard | Increasingly |
| Pattern Lock | Weak–Mod | Easy | Easy | Yes |
Practical, Ethical & Real-World Considerations
- Balance convenience vs protection: stronger controls can reduce user adoption; education crucial.
- Regulatory landscape (GDPR, CCPA) mandates explicit consent for PII collection & breach notification.
- Developers should embrace privacy-by-design: data minimisation, end-to-end encryption, secure coding.
- Societal impact: language & cognitive effects of pervasive texting; digital divide amplified by security literacy gap.
- Corporate responsibility: provide secure defaults (e.g., Android’s Scoped Storage, iOS Secure Enclave).
- Future trends: 5G slicing implies new isolation challenges; IoT integration broadens mobile attack vectors.
End of consolidated study notes for “Security in Mobile Computing – Topic 6: Device Security”.