Instiq
Chapter 1 · Standards and Concepts·v1.0.0·Updated 7/17/2026·~13 min

What's changed: Initial version

1.4Cloud and on-premises

Key points

Covers the difference between on-premises (devices in your own facility) and cloud (using a provider's infrastructure), the cloud service models SaaS/PaaS/IaaS, the deployment models public/private/hybrid cloud, and the spread of remote work and hybrid work, as basic judgment about "which model fits a requirement."

Today's networks do not stay inside one's own building—email, files, and business apps live in the cloud, and employees access them from home or on the move. To grasp a user's trouble correctly, support staff must distinguish "is that service on devices we placed ourselves (on-premises) or in a provider's cloud?" and "which service model (SaaS/PaaS/IaaS)?" This section organizes the difference and trade-offs between cloud and on-prem, and the shift to a world where remote work is normal, as introductory-level judgment.

1.4.1On-premises vs. cloud

  • On-premises means installing servers and devices within your own facilities and maintaining/operating them yourself. Upfront cost is high, but devices and data stay fully under your control (suited to confidentiality and customization needs). You also handle failures and scaling yourselves.
  • Cloud means using a provider's infrastructure on a pay-as-you-go basis. You can scale up or down quickly on demand (scalability) and reduce the burden of buying and maintaining hardware. Network-wise, because "the service is outside your own site," reachability over the internet and authentication become prerequisites.

1.4.2SaaS/PaaS/IaaS and deployment models

  • Service models come in three levels by "how much the provider handles": SaaS = use a finished app as-is (e.g., webmail or office suites; the user only configures). PaaS = rent a platform (OS and middleware) to develop and run apps (for developers). IaaS = rent only the infrastructure (virtual servers, networking), building everything above the OS yourself.
  • Classification by deployment: public cloud = a provider's shared cloud offered to the general public. private cloud = a cloud dedicated to one organization (for control and compliance). hybrid cloud = combining on-prem/private with public and using each where it fits (sensitive data on private, variable load on public, etc.).
  • Remote work (working from home or elsewhere) and hybrid work (mixing office and home) have spread, so employees access the cloud and internal systems from anywhere. Secure connections use VPNs and the like (details in a later chapter).
Exam point

Most-tested: on-prem = self-installed and self-operated (strong control, high upfront cost) vs. cloud = a provider's infrastructure on pay-as-you-go (strong scalability, low ops burden); SaaS = finished app, PaaS = development platform, IaaS = infrastructure only; public/private/hybrid differ by deployment scope. Be able to choose which model fits which requirement.

Suppose you handle IT support for a small company and the owner asks, "I want to introduce a new expense-reimbursement app, with as little operational effort as possible." Reflexively answering "let us buy a server and build it on-premises" is an overkill judgment against the requirement (minimal effort), because you would shoulder server procurement, OS maintenance, backups, and failure response yourselves—heavy for a small company. If the requirement is "use a finished business app right away and leave operations to someone else," the natural answer is SaaS (a web-based expense service), letting users focus on configuration and data entry. If instead the requirement is "a platform to run our own custom app, managing OS and middleware ourselves," PaaS is the candidate; "rent only virtual servers and networking and assemble everything ourselves" points to IaaS. Further, if a constraint says "only customer confidential data must not leave our premises," keeping the sensitive part on private (or on-prem) while using public for other variable processing—a hybrid—is realistic. And if employees will use it from home (remote work), secure access from outside must be part of the design. The key is the judgment pattern: rather than choosing by image ("cloud because it is new," "on-prem because it feels safe"), choose and combine SaaS/PaaS/IaaS and public/private/hybrid from the requirements—operational effort, control/confidentiality, scalability, and working style. At the introductory stage, holding this measuring stick of choosing a model from requirements matters more than detailed build steps.

ModelWhat the provider managesFits when
On-premises(You manage everything)Full control, confidentiality, custom needs
IaaSUp to infrastructure (VMs, networking)Rent infrastructure, build the rest yourself
PaaSUp to OS and middlewareWant a platform to develop and run apps
SaaSThe finished app itselfUse a finished app now, leave ops to the provider
Warning

Trap: "Cloud is always cheaper and safer than on-prem, so even confidential data must go to the public cloud" is wrong—the choice depends on requirements, and when control or confidentiality is paramount, on-premises or a private cloud can be appropriate. Also wrong: "SaaS and IaaS are both cloud, so the user's effort is the same"—SaaS is a finished app with little effort, while IaaS requires you to build everything above the OS, a much larger effort.

On-premises vs. cloud, SaaS/PaaS/IaaS, deployment models, remote work.
Judging which model fits a requirement

1.4.3Section summary

  • On-premises is self-installed with strong control (high upfront cost); cloud uses a provider's infrastructure on pay-as-you-go with high scalability (low ops burden)
  • Service models split into SaaS (finished app) / PaaS (platform) / IaaS (infrastructure only); deployment into public/private/hybrid
  • Remote work/hybrid work have spread, making secure access from outside to cloud and internal systems a design premise

Sign in to track progress — Log in.

Quick check

(just a quick review)

Q1. A small-company owner says, "I want to use a finished expense app right away and take on as little server operation and OS maintenance as possible." Which service model fits best?

Q2. A requirement states: "customer confidential data must not leave our premises, but seasonally spiking aggregation needs flexible scalability." Which architectural approach fits best?

Q3. When explaining the difference between on-premises and cloud to a new hire, which statement is most accurate?

Check your understandingPractice questions for Chapter 1: Standards and Concepts