How unit 2 is examined
This unit covers the computing models behind the cloud (utility, elastic, grid, HPC), the hypervisor, virtualization benefits and pitfalls, and multitenancy; the hypervisor carries the most marks, followed by virtualization benefits, pitfalls and multitenancy.
Utility Computing
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>Utility computing is a service model in which computing resources such as processing, storage and software are provided on demand and billed by metered usage, like electricity or water.</mark>
Key points.
- The customer pays only for what is consumed, so there is no large upfront investment in hardware.
- Usage is metered (CPU hours, GB stored, GB transferred) and charged at a rate per unit, so total cost = units used x rate.
- Resources are shared by many customers and supplied on demand, which lets the provider reach economies of scale.
- It is the business model behind cloud computing; Amazon EC2 billed per hour or second is the standard example.
Elastic computing
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>Elastic computing is the ability of a cloud to automatically add resources when demand rises and release them when demand falls, so capacity always matches the workload.</mark>
Key points.
- Scale-out (horizontal) adds more instances, while scale-up (vertical) gives one instance more CPU or RAM.
- Auto-scaling rules watch a metric such as CPU above 70% and add or remove instances without human action.
- Elasticity avoids both over-provisioning (wasted cost) and under-provisioning (slow service or crashes).
- Elasticity is short-term and automatic, whereas scalability is the long-term ability of a system to grow; an example is an e-commerce site scaling out during a festival sale and scaling in afterwards.
Grid computing
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>Grid computing links geographically distributed, heterogeneous computers under different administrations into one virtual supercomputer that solves a large task by sharing their idle capacity.</mark>
Key points.
- A large job is split into parts that run on many loosely coupled nodes, and middleware such as the Globus Toolkit handles scheduling, security and resource discovery.
- Nodes may differ in hardware and operating system and belong to different organisations.
- It suits batch, compute-intensive science such as SETI@home, weather modelling and protein folding.
- Unlike the cloud, a grid is mostly job-oriented and shared by collaboration, without pay-per-use billing or on-demand self-service.
Study of Hypervisor
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">High weight</span>
Definition. <mark>A hypervisor (virtual machine monitor) is software or firmware that creates and runs virtual machines by sharing one physical machine's CPU, memory, storage and network among many isolated guest operating systems.</mark>
Diagram. <figure class="ds-fig" style="margin:1.4rem 0;overflow-x:auto"><svg xmlns="http://www.w3.org/2000/svg" id="dsfig-u2-01" viewBox="0 0 381 338" width="381" height="338" role="img" aria-label="Left, Type-1 (bare metal): VMs on hypervisor (H1) directly on hardware (HW). Right, Type-2 (hosted): hardware (HW2), host OS, hypervisor (H2), then VM (VC)."><style>#dsfig-u2-01 .e{stroke:#454C5A;stroke-width:1.4;fill:none}#dsfig-u2-01 .e.hi{stroke:#2340B8;stroke-width:2.6}#dsfig-u2-01 .n{fill:#FFFFFF;stroke:#16181D;stroke-width:1.4}#dsfig-u2-01 .n.hi{fill:#E3E9FC;stroke:#2340B8;stroke-width:2.2}#dsfig-u2-01 .n.rb-b{fill:#16181D;stroke:#16181D}#dsfig-u2-01 .n.rb-r{fill:#BD3227;stroke:#BD3227}#dsfig-u2-01 text{font-family:"JetBrains Mono",ui-monospace,Menlo,Consolas,monospace;font-size:13px}#dsfig-u2-01 .t{fill:#16181D;font-weight:500}#dsfig-u2-01 .t.inv{fill:#FFFFFF;font-weight:700}#dsfig-u2-01 .kd{stroke:#16181D;stroke-width:1.2}#dsfig-u2-01 .dot{fill:#16181D}#dsfig-u2-01 .ann{fill:#2340B8;font-size:11px;font-weight:700}#dsfig-u2-01 .lbl{fill:#6F7787;font-family:system-ui,-apple-system,sans-serif;font-size:12px;font-weight:700}#dsfig-u2-01 .ptr{fill:#2340B8;font-size:12px;font-weight:700}#dsfig-u2-01 .ah{fill:#454C5A}#dsfig-u2-01 .ah.hi{fill:#2340B8}#dsfig-u2-01 .wl rect{fill:#FFFFFF;stroke:#DCE0E7}#dsfig-u2-01 .wl .t{font-size:12px;font-weight:700}#dsfig-u2-01 .wl.hi rect{fill:#2340B8;stroke:#2340B8}#dsfig-u2-01 .wl.hi .t{fill:#FFFFFF}html.dark #dsfig-u2-01 .e{stroke:#B1B7C3}html.dark #dsfig-u2-01 .e.hi{stroke:#8FA3FF}html.dark #dsfig-u2-01 .n{fill:#161920;stroke:#E6E8ED}html.dark #dsfig-u2-01 .n.hi{fill:#1E2748;stroke:#8FA3FF}html.dark #dsfig-u2-01 .n.rb-b{fill:#E6E8ED;stroke:#E6E8ED}html.dark #dsfig-u2-01 .n.rb-r{fill:#FF7E71;stroke:#FF7E71}html.dark #dsfig-u2-01 .t{fill:#E6E8ED}html.dark #dsfig-u2-01 .t.inv{fill:#0F1115}html.dark #dsfig-u2-01 .kd{stroke:#E6E8ED}html.dark #dsfig-u2-01 .dot{fill:#E6E8ED}html.dark #dsfig-u2-01 .ann{fill:#8FA3FF}html.dark #dsfig-u2-01 .lbl{fill:#858D9C}html.dark #dsfig-u2-01 .ptr{fill:#8FA3FF}html.dark #dsfig-u2-01 .ah{fill:#B1B7C3}html.dark #dsfig-u2-01 .ah.hi{fill:#8FA3FF}html.dark #dsfig-u2-01 .wl rect{fill:#161920;stroke:#2A2E37}html.dark #dsfig-u2-01 .wl.hi rect{fill:#8FA3FF;stroke:#8FA3FF}html.dark #dsfig-u2-01 .wl.hi .t{fill:#0F1115}</style><defs><marker id="ah2" viewBox="0 0 10 10" refX="9" refY="5" markerWidth="7" markerHeight="7" orient="auto-start-reverse"><path class="ah" d="M0,1 L9,5 L0,9 z"/></marker><marker id="ahh2" viewBox="0 0 10 10" refX="9" refY="5" markerWidth="7" markerHeight="7" orient="auto-start-reverse"><path class="ah hi" d="M0,1 L9,5 L0,9 z"/></marker></defs><path class="e" d="M83,231 L83,277" marker-end="url(#ah2)"/><path class="e" d="M48.5,143 L73.6,193.2" marker-end="url(#ah2)"/><path class="e" d="M117.5,143 L92.4,193.2" marker-end="url(#ah2)"/><path class="e" d="M341,231 L341,277" marker-end="url(#ah2)"/><path class="e" d="M341,145 L341,191" marker-end="url(#ah2)"/><path class="e" d="M341,59 L341,105" marker-end="url(#ah2)"/><circle class="n" cx="83" cy="298" r="18"/><text class="t" x="83" y="298" dy=".35em" text-anchor="middle">HW</text><circle class="n" cx="83" cy="212" r="18"/><text class="t" x="83" y="212" dy=".35em" text-anchor="middle">H1</text><circle class="n" cx="40" cy="126" r="18"/><text class="t" x="40" y="126" dy=".35em" text-anchor="middle">VA</text><circle class="n" cx="126" cy="126" r="18"/><text class="t" x="126" y="126" dy=".35em" text-anchor="middle">VB</text><circle class="n" cx="341" cy="298" r="18"/><text class="t" x="341" y="298" dy=".35em" text-anchor="middle">HW2</text><circle class="n" cx="341" cy="212" r="18"/><text class="t" x="341" y="212" dy=".35em" text-anchor="middle">OS</text><circle class="n" cx="341" cy="126" r="18"/><text class="t" x="341" y="126" dy=".35em" text-anchor="middle">H2</text><circle class="n" cx="341" cy="40" r="18"/><text class="t" x="341" y="40" dy=".35em" text-anchor="middle">VC</text></svg><figcaption style="font-size:.82em;opacity:.72;margin-top:.45rem">Left, Type-1 (bare metal): VMs on hypervisor (H1) directly on hardware (HW). Right, Type-2 (hosted): hardware (HW2), host OS, hypervisor (H2), then VM (VC).</figcaption></figure>
Key points.
- The hypervisor sits between hardware and guest operating systems and gives each virtual machine its own virtual CPU, memory, disk and network card.
- A Type-1 hypervisor installs directly on the physical hardware with no host operating system beneath it, and examples are VMware ESXi, Microsoft Hyper-V, Xen and KVM.
- A Type-2 hypervisor runs as an application on a host operating system, and examples are VirtualBox and VMware Workstation.
- Type-1 is called bare metal (native) because it runs on the bare hardware itself, controlling resources directly and needing no host OS layer.
- Because there is no extra OS layer, Type-1 has lower overhead, better performance and a smaller attack surface, so data centres and clouds use it.
- Type-2 is easier to install and suits desktops and testing, but every request passes through the host OS, so it is slower and depends on the host's security.
- The hypervisor isolates VMs so that a crash or attack in one does not affect the others, and it enables live migration and snapshots.
| Basis | Type-1 (bare metal) | Type-2 (hosted) |
|---|---|---|
| Runs on | Hardware directly | Host OS |
| Performance | High | Lower |
| Security | Smaller attack surface | Depends on host OS |
| Use | Data centres, cloud | Desktop, testing |
| Examples | ESXi, Hyper-V, Xen | VirtualBox, Workstation |
Answer frame. For Q4 open by defining a Type-1 hypervisor as one installed directly on hardware; draw the left half of the diagram; develop points 2, 4, 5 and 7; close that no host OS lies between hypervisor and hardware, hence the name bare metal. For the short-note question (Jun 2025) write a definition, the two types with examples, and one use in cloud: the hypervisor lets a provider run many tenants' VMs on one server. The other options in that question are answered in Unit 1 (PaaS), the Elastic computing section here, and Unit 4 (QoS, VM security).
Asked: [7 marks] (Dec 2024) Why type-1 hypervisor is called as bare metal architecture? Asked: [14 marks] (Jun 2025) Write short notes (any three): a) Hypervisor b) PaaS c) Elastic Computing d) QOS issues in Cloud Computing e) VM Security
Virtualization applications in enterprises
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Low weight</span>
Definition. <mark>Virtualization in the cloud is the creation of virtual versions of servers, storage and networks on shared physical hardware, so many isolated workloads run on one machine.</mark>
Key points.
- Resource utilisation rises because many VMs share one server, which lifts typical use from about 10-15% to 60-80%.
- Isolation keeps each VM and tenant separate, so one failure or attack does not spread.
- Scalability improves because new VMs are created in minutes, and cost falls through fewer servers, less power and less space.
- Live migration moves running VMs between hosts with no downtime, which allows maintenance without interruption.
- Disaster recovery is simpler because VM snapshots and images are copied to another site and restarted quickly.
- Management is centralised, since one console provisions, monitors and backs up all VMs; enterprises also use it for server consolidation, test environments and virtual desktops.
Asked: [7 marks] (Dec 2024) What are the benefits of virtualization in cloud?
High-performance computing
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>High-performance computing (HPC) aggregates many processors, often as a cluster of tightly connected nodes, to solve very large problems far faster than a single computer.</mark>
Key points.
- Nodes are linked by fast interconnects and run tasks in parallel, with speed measured in FLOPS.
- Typical uses are weather forecasting, genome analysis, engineering simulation and AI training.
- The cloud offers HPC on demand, so users rent thousands of cores by the hour instead of buying a cluster.
- Unlike a grid, an HPC cluster is tightly coupled and homogeneous in one location.
Pitfalls of virtualization
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Low weight</span>
Definition. <mark>Pitfalls of virtualization are the risks and costs that come with running many VMs on shared hardware in an enterprise cloud, such as sprawl, overhead, security and single points of failure.</mark>
Key points.
- VM sprawl: VMs are so easy to create that unused ones pile up, wasting resources and complicating management.
- Performance overhead: the hypervisor uses CPU and memory, and VMs on one host compete for I/O, causing the noisy-neighbour effect.
- Security: a compromised hypervisor or a VM escape exposes every VM on that host, and shared resources open side channels.
- Licensing: software licences may be priced per core or per VM, so costs and compliance become complex.
- Single point of failure: if a host fails, all VMs on it go down together.
- Mitigation: use governance and lifecycle policies to remove idle VMs, monitor performance, plan capacity, harden and patch the hypervisor, and run hosts in clusters with failover.
Asked: [7 marks] (Jun 2025) What are the common pitfalls of virtualization in enterprise cloud computing? How can they be mitigated?
Multitenant software
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Low weight</span>
Definition. <mark>Multitenancy is an architecture in which one running instance of a software application serves many customers (tenants), each with logically isolated data and configuration.</mark>
Key points.
- A tenant is a customer organisation; all tenants share the same code, servers and database platform.
- Each tenant sees only its own data and can customise branding and settings, so it feels like a private application.
- It lowers cost, since one instance is upgraded and maintained for everyone; Salesforce and Gmail for business are examples.
- Its main risks are data leakage between tenants and noisy neighbours, so isolation is the key design issue.
- Tenant data is separated either by a tenant ID in shared tables (multi-entity support) or by a separate schema per tenant (multi-schema approach).
| Basis | Multi-entity (shared schema) | Multi-schema (separate schema) |
|---|---|---|
| Data layout | All tenants in the same tables, with a tenant ID column | Each tenant has its own schema or tables |
| Isolation | Weaker, depends on correct filtering | Stronger, at database level |
| Scalability | Best, thousands of tenants | Limited by number of schemas |
| Customisation | Hard, all share one structure | Easy, per-tenant changes |
| Cost | Lowest | Higher |
Answer frame. Open with the definition of multitenancy; draw one instance with tenants A, B, C beneath it; develop points 1-4; then give the comparison table row by row; close that shared schema suits many small tenants while separate schema suits tenants needing isolation.
Asked: [7 marks] (Jun 2025) Explain the concept of multitenant software and compare multi-entity support with the multi-schema approach.
Multi-entity support
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>Multi-entity support keeps all tenants in one shared database and schema, and separates their rows using a tenant (entity) ID column carried in every table.</mark>
Key points.
- Every query filters by tenant ID, so each tenant sees only its own rows.
- It is the cheapest and most scalable model because there is one schema to manage and upgrade.
- Its weakness is that a missing filter can leak data, and all tenants must use the same structure.
- It suits SaaS products with very many small tenants.
Multi schema approach
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>In the multi-schema approach all tenants share one database server, but each tenant has its own separate schema (set of tables).</mark>
Key points.
- Tenant data is separated by the database itself, so isolation is stronger than with a tenant ID column.
- Each tenant can have custom fields or tables, and backup and restore can be done per tenant.
- It costs more to run, and schema changes must be applied to every tenant's schema.
- It suits fewer, larger tenants that need stronger isolation.
Last-minute revision
- Utility computing is pay-per-use metered service: cost = units used x rate.
- Elasticity is automatic scale-out and scale-in to match demand; scalability is long-term growth.
- Grid is loosely coupled and heterogeneous across organisations (Globus); HPC is a tightly coupled cluster.
- A hypervisor (VMM) creates and manages VMs on shared hardware.
- Type-1 is bare metal and runs on hardware (ESXi, Hyper-V, Xen, KVM); Type-2 is hosted on an OS (VirtualBox, Workstation).
- Type-1 is bare metal because it has no host OS layer and controls hardware directly.
- Benefits of virtualization: utilisation, isolation, scalability, cost, live migration, disaster recovery, central management.
- Pitfalls: sprawl, overhead, security, licensing and single point of failure; fix with governance, monitoring, capacity planning and hardening.
- Multitenancy is one instance serving many tenants with isolated data.
- Multi-entity uses a shared schema with a tenant ID; multi-schema uses one schema per tenant.
Memory hooks
- Type-1 = "naked" on hardware (bare metal); Type-2 = "wearing" a host OS coat.
- Pitfalls = S-O-S-L-S: Sprawl, Overhead, Security, Licensing, Single point of failure.
- Utility = electricity meter; Elastic = rubber band; Grid = volunteers' idle PCs.
- Multi-entity = one flat, name tags on rooms; multi-schema = separate flats in one building.
Coverage checklist
- Utility Computing: definition, metered pay-per-use, key points.
- Elastic computing: definition, scaling types, key points (also Jun 2025 short-note option c).
- Grid computing: definition, Globus, key points.
- Study of Hypervisor: types, diagram, table, bare metal (Dec 2024, 7 marks); short notes hypervisor (Jun 2025, 14 marks).
- Virtualization applications in enterprises: benefits of virtualization (Dec 2024, 7 marks).
- High-performance computing: definition, key points.
- Pitfalls of virtualization: pitfalls and mitigation (Jun 2025, 7 marks).
- Multitenant software: concept and comparison (Jun 2025, 7 marks).
- Multi-entity support: definition, key points.
- Multi schema approach: definition, key points.