Mobile Technology, Cloud Computing and IoT

This chapter closes Unit 9 with two sub-items beyond the wired protocol stack of the course's Computer Networks chapters: Mobile Technology (GSM and CDMA, mobile computing architecture, mobile IP, wireless LANs and topologies, GPRS and SMS) and Cloud Computing and IoT (service and deployment models, virtualisation, the basics of IoT).

GSM and CDMA, and Mobile Computing Architecture

GSM (Global System for Mobile Communications) divides available spectrum by TIME AND FREQUENCY (TDMA/FDMA combined) — each user gets a specific frequency channel AND a specific time slot within it, so multiple users literally take turns on the same frequency. CDMA (Code Division Multiple Access) takes a fundamentally different approach: EVERY user transmits over the SAME entire frequency band and at the SAME time, but each is assigned a unique, near-orthogonal spreading CODE — the receiver can extract one user's signal from the combined mixture by correlating against that user's own code, which is what lets many users share the same spectrum simultaneously without the explicit time/frequency partitioning GSM needs.

  • Mobile computing architecture: typically layered as fixed hosts and servers on the wired network, a MOBILE SUPPORT STATION (base station) bridging wired and wireless, and mobile hosts communicating over the wireless link — the mobile host's location changes over time, which is precisely the complication wired network architecture never has to handle.
  • Mobile IP solves exactly this: a mobile host keeps its permanent HOME ADDRESS regardless of location, registers a temporary CARE-OF ADDRESS with a HOME AGENT whenever it moves to a new (foreign) network, and the home agent then TUNNELS (encapsulates and forwards) incoming packets addressed to the home address onward to the current care-of address — transparently to whoever is sending to the mobile host, who never needs to know the host has moved at all.

Wireless Networks and Topologies, GPRS and SMS

  • Wireless LANs most commonly use INFRASTRUCTURE mode (all devices communicate through a central access point, which also bridges to the wired network) or AD HOC mode (devices communicate directly, peer-to-peer, with no central access point at all) — the syllabus separately names MOBILE AD HOC NETWORKS (MANETs), where every node can also act as a router forwarding others' traffic, and the network's topology itself changes as nodes move.
  • Cellular topology: geographic area divided into hexagonal CELLS, each served by its own base station on a limited set of frequencies; the same frequencies are REUSED in non-adjacent cells (frequency reuse) — this is exactly why a huge number of users can be served with a small, fixed spectrum allocation, at the cost of a HANDOFF (handover) mechanism needed to keep a call connected as a moving user crosses from one cell into another.
  • GPRS (General Packet Radio Service) adds PACKET-SWITCHED data service on top of the (originally circuit-switched) GSM voice network — users are billed by data VOLUME rather than connection TIME, and the connection is 'always on' rather than dialled per session, a genuine architectural shift from voice-era GSM. SMS (Short Message Service) sends short text messages over the network's SIGNALLING channel rather than its voice/data traffic channel, which is exactly why an SMS can still be delivered even while a voice call occupies the same phone's traffic channel.

Cloud Computing: Service and Deployment Models

The three cloud service models, by how much the provider manages
ModelCustomer managesExample
IaaS (Infrastructure)The OS, runtime, applications and data — provider supplies only virtualized hardware (compute, storage, network)A rented virtual machine
PaaS (Platform)Only the application and data — provider also manages the OS and runtime environmentA managed application-hosting platform
SaaS (Software)Only their own data and configuration — the provider manages everything else, delivering a ready-to-use applicationA web-based email or office-suite service

Public cloud infrastructure is owned by a third-party provider and shared across many customers over the internet (economical, less direct control); private cloud infrastructure is dedicated to a single organisation (more control and customisation, at higher cost); a hybrid cloud combines both, letting workloads move between them as needed. Underlying all three service models is virtualization (covered in Unit 5) and shared resource management governed by a Service Level Agreement (SLA) — the contractual commitment on uptime, performance and support that the customer actually relies on.

Basics of the Internet of Things (IoT)

The Internet of Things connects physical objects — sensors, actuators, everyday devices — to a network so they can sense their environment, communicate, and often act on it, largely without direct human operation at each step. A typical IoT stack has a PERCEPTION layer (sensors/actuators gathering physical data), a NETWORK layer (transmitting that data, often over low-power wireless protocols distinct from conventional Wi-Fi because IoT devices are commonly battery-constrained), and an APPLICATION layer (processing and acting on the data, frequently in the cloud given the low computational power of the devices themselves) — the same three-layer shape recurring across most IoT architectures regardless of the specific application (smart homes, industrial monitoring, wearables).

Key takeaways

  • GSM partitions by time and frequency (TDMA/FDMA); CDMA lets every user share the same time and frequency using unique spreading codes.
  • Mobile IP tunnels packets from a fixed home address to a mobile host's current care-of address via a home agent, transparently to the sender.
  • Cellular topology reuses frequencies across non-adjacent cells, needing a handoff mechanism as a user crosses cell boundaries; GPRS adds packet-switched, always-on data over GSM; SMS rides the signalling channel.
  • IaaS/PaaS/SaaS differ in how much the customer versus the provider manages; public/private/hybrid describe ownership and sharing of the underlying infrastructure.
  • A typical IoT stack has perception, network and application layers, with low-power protocols and cloud-based processing common across most IoT architectures.

Practice questions (8)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

  1. CDMA, compared to GSM, allows multiple users to share spectrum by:

    1. Assigning each user a unique spreading code while all transmit on the same frequency and time
    2. Giving each user a completely separate physical wire
    3. Allowing only one user on the network at any time
    4. Requiring no receiver at all
    Show answer

    Answer: A — Assigning each user a unique spreading code while all transmit on the same frequency and time

    Every CDMA user transmits over the same band and time simultaneously; the receiver separates one user's signal from the combined mixture by correlating against that user's own near-orthogonal code, unlike GSM's explicit time/frequency partitioning.
  2. In Mobile IP, when a mobile host moves to a foreign network, it registers a temporary address called the:

    1. Care-of address
    2. Home address
    3. MAC address
    4. Loopback address
    Show answer

    Answer: A — Care-of address

    The permanent home address never changes; the temporary care-of address, registered with the home agent, is what lets packets addressed to the home address be tunnelled onward to wherever the host currently is.
  3. In a cellular network, the same set of frequencies being reused in non-adjacent cells is called:

    1. Frequency reuse
    2. Handoff
    3. Tunnelling
    4. Multiplexing only
    Show answer

    Answer: A — Frequency reuse

    Frequency reuse is exactly what lets a small, fixed spectrum allocation serve a huge number of users across a large geographic area, at the cost of needing a handoff mechanism as a user's call crosses cell boundaries.
  4. In the IaaS/PaaS/SaaS cloud service models, moving from IaaS to SaaS means:

    1. The customer manages progressively less, and the provider manages progressively more
    2. The customer manages progressively more of the underlying hardware
    3. No provider is involved at all in SaaS
    4. There is no difference between the three models
    Show answer

    Answer: A — The customer manages progressively less, and the provider manages progressively more

    IaaS gives the customer the most to manage (OS upward) while the provider handles only virtualized hardware; SaaS is the opposite end, where the provider manages everything and the customer manages only their own data/configuration.
  5. A typical IoT architecture's perception layer is responsible for:

    1. Sensors and actuators gathering physical data from the environment
    2. Processing all data exclusively in the cloud
    3. Transmitting data over low-power wireless protocols
    4. Billing the customer for cloud usage
    Show answer

    Answer: A — Sensors and actuators gathering physical data from the environment

    The perception layer is where physical sensing/actuation happens; transmitting the gathered data is the network layer's job, and processing/acting on it is the application layer's — three distinct responsibilities in the standard stack.
  6. Which of the following are true of GPRS compared to the original circuit-switched GSM voice network? (More than one option may be correct.)

    1. It adds packet-switched data service
    2. Users are billed by data volume rather than connection time
    3. The connection is 'always on' rather than dialled per session
    4. It removes voice calling entirely from the network
    Show answer

    Answer: A — It adds packet-switched data service; B — Users are billed by data volume rather than connection time; C — The connection is 'always on' rather than dialled per session

    GPRS adds a packet-switched, always-on, volume-billed data layer alongside GSM's original voice service — it augments rather than replaces voice calling.
  7. Name the contractual document that specifies a cloud provider's committed uptime, performance and support levels to a customer.

    Numerical answer — type the value.

    Show answer

    Answer: SLA

    The Service Level Agreement (SLA) is the concrete commitment a customer can actually hold a provider to, distinct from the marketing description of a cloud service.
  8. A private cloud, compared to a public cloud, is best described as:

    1. Infrastructure dedicated to a single organisation, offering more control at higher cost
    2. Infrastructure shared across many unrelated customers over the internet
    3. Always cheaper than a public cloud
    4. Identical to a hybrid cloud in every respect
    Show answer

    Answer: A — Infrastructure dedicated to a single organisation, offering more control at higher cost

    A private cloud's infrastructure serves one organisation exclusively, trading the public cloud's economy of shared infrastructure for greater control and customisation, typically at higher cost.