Cloud Concepts
The five cloud "-ilities"
"Can grow" and "grows and shrinks itself" sound like the same thing, and the Cloud Digital Leader exam keeps testing whether you know they are not: a scalable system can add capacity, while an elastic one does it automatically in both directions. That pair sits inside five words for why cloud is valuable (scalability, elasticity, reliability, flexibility, agility), and telling them apart is what lets you pick the one a scenario is describing. The diagram above groups the five by whether they describe the system or the organization.
Scalability is the ability to change a system's capacity to match demand: add more machines (or bigger machines) when load grows, remove them when it shrinks. Elasticity is scalability done automatically and quickly, in both directions: capacity expands as traffic rises and contracts as it falls, without a human placing a hardware order. Every elastic system is scalable, but a system can be scalable (you can add capacity) without being elastic (it does not do so on its own). When a question contrasts "can grow" with "grows and shrinks itself," it is testing exactly this distinction.
Reliability is the system's ability to keep performing its function and to recover when something fails. In cloud terms this is achieved by running across independent failure domains (see Regions and zones below) so one failure does not take everything down.
Flexibility is the freedom to choose, combine, and change technologies, regions, and even providers without buying and installing new hardware first. Agility is the business speed that flexibility unlocks: teams ship and change products faster because they are no longer waiting weeks for servers to arrive. Google frames the cloud's benefits around scalable, flexible, agile, secure, cost-effective, and strategic outcomes; the Google Cloud Architecture Framework[1] organizes its guidance around these same operational and reliability properties.
A useful mental model: the first three (scalability, elasticity, reliability) are properties of the system, while flexibility and agility are properties of the organization the system enables. The exam rewards matching the scenario to the right layer: "we want capacity to follow traffic automatically" is elasticity; "we can launch a new feature in days, not quarters" is agility.
TCO and the CapEx-to-OpEx shift
Cloud changes what you pay for and when: CapEx, OpEx, and TCO are the three terms that capture the shift, alongside the three pricing levers the exam expects you to know. The diagram above contrasts the two spending models the shift moves between.
Capital expenditure (CapEx) is money spent up front on assets you own and depreciate over years: in traditional IT, that is buying servers, storage, and networking gear sized for your expected peak demand. Operating expenditure (OpEx) is money spent on ongoing consumption. Google states that "in the cloud, the costs for most cloud resources are treated as OpEx, where costs are incurred when the cloud resources are consumed": you pay for what you use, when you use it, with no up-front hardware purchase. This is the CapEx-to-OpEx shift, and it is the single most-tested cost idea on the exam.
Total cost of ownership (TCO) is the complete cost of running a workload, not just the price tag of the hardware. On-premises, TCO hides many costs: power, cooling, data-center real estate, hardware refresh every few years, and the staff to run it all. Because the cloud folds those into a single usage-based bill and removes the over-provisioning needed to cover peak demand, it typically lowers TCO, which is why "reduce TCO" is a common correct answer when a scenario complains about idle, over-bought hardware.
Three pricing levers reinforce the OpEx model, and the exam expects the concepts, not dollar figures:
| Lever | What it is | Commitment? |
|---|---|---|
| Pay-as-you-go | Default billing (pay only for resources consumed) | None |
| Sustained-use discounts | Automatic discount that grows the longer a resource runs in a month | None (applied automatically) |
| Committed-use discounts | Discounted price in exchange for committing to a 1- or 3-year term | Yes (a 1- or 3-year commitment) |
The trap here is mixing up the two discounts. A committed use discount[2] requires you to commit to a one- or three-year term; a sustained use discount is automatic and needs no commitment. "We want a lower rate but cannot commit to a term" points to sustained-use (or pay-as-you-go); "we will run this steadily for years and want the best rate" points to committed-use. Cost discipline around these levers is the subject of the cost-optimization pillar[3] of the Architecture Framework.
Choosing a deployment model
Building on the cost and "-ility" vocabulary above, the key idea (stated once and reused) is that the deployment model is chosen per workload, driven by where the data must live and how much control the workload needs, not by picking one model for the whole company. Public, private, hybrid, and multicloud are the four choices a scenario asks you to pick between and justify.
- Public cloud runs your workloads on a provider's shared, on-demand infrastructure (Google Cloud, AWS, Azure). It maximizes the OpEx model, elasticity, and agility, and usually delivers the lowest TCO, at the cost of less low-level control of the underlying hardware.
- Private cloud dedicates infrastructure to a single organization, often on-premises. It is chosen for data-residency rules, regulatory control, or legacy systems that cannot move. The trade-off is mostly CapEx and limited elasticity: you still own and size the hardware.
- Hybrid cloud connects a private or on-premises environment with a public cloud so workloads can span both. It is the answer when some of the estate is constrained (a regulated database stays private) while the rest goes cloud-native. The cost is network and integration complexity between the two sides.
- Multicloud uses two or more public-cloud providers at once. It is justified by a concrete need, avoiding vendor lock-in, or using a best-of-breed service that only one provider offers, and carries the most integration and skills overhead.
The exam's recurring trap is treating "use everything" as automatically safest. Multicloud "just to feel safer" is a wrong answer: each added provider multiplies integration and operational cost, so it must be tied to a stated business reason. Likewise, when a scenario says data legally cannot leave a country and no in-region option exists, a public-cloud-only design may not fit. A private or hybrid model is the conceptual answer.
Network terms: regions, zones, and latency
Regions versus zones, plus the everyday network terms the exam lists, are the geographic and performance building blocks of cloud, and each is defined at first use here, no networking background required.
Google defines a region as an "independent geographic area that consist of zones," and a zone as a "deployment area for Google Cloud resources within a region." The relationship is the load-bearing fact: a region contains three or more zones, each in a separate physical data center, and (quoting Google) "zones should be considered a single failure domain within a region." The diagram above shows this nesting.
Why it matters for design:
- Place resources in a region close to your users to reduce latency, the time it takes a packet to travel from source to destination. Lower latency means a more responsive application.
- Spread resources across multiple zones so that a single-zone failure (a power or hardware fault in one data center) does not take the whole workload down. This is how the reliability property from the first section is actually achieved.
- Multi-region services (such as Cloud Storage and Spanner) replicate data across regions and are designed to keep working after the loss of an entire region, per Google's geography and regions[4] documentation.
The exam also lists supporting terms you should simply recognize: an IP address is the numeric address that identifies a device on a network; DNS (Domain Name System) translates human-readable names like example.com into IP addresses; an ISP (Internet Service Provider) is the company that connects you to the internet; bandwidth is how much data can move per second (capacity), as distinct from latency (delay). Bandwidth and latency together describe network performance: high bandwidth with high latency still feels slow for interactive use.
Google's global infrastructure and network
What Google owns is what makes its cloud globally fast, and why traffic on Google Cloud largely avoids the public internet, all at a leadership altitude with no hands-on networking required. The diagram above traces a request along Google's backbone next to the public-internet path it replaces.
Google operates a global, privately owned fiber-optic network (including subsea cables) that connects its regions to each other, plus a worldwide set of edge points of presence (PoPs). Google's documentation describes "a global network of peering points of presence, which means that customer traffic can travel within the Google network until it's close to its destination," and notes connections "from over 200 locations" worldwide, per the network edge locations[5] reference.
The single idea to carry into the exam: because Google owns this backbone end to end, user traffic enters Google's network near the user and stays on Google's private fiber for most of its journey, rather than crossing the open public internet hop by hop. The business outcomes are lower latency, higher and more consistent bandwidth, and improved security, and crucially, the customer gets that global reach without building any of the network themselves. Google's global backbone "provides tremendous flexibility for load-balancing, and reduces end-user latency,"[4] which is the conceptual link back to the scalability and reliability properties from the first section.
For a Cloud Digital Leader, the depth stops here: you are expected to know that Google designs, owns, and runs this global infrastructure and why that benefits customers, not how to configure routing, network tiers, or load balancers, which belong to the associate-level networking exams.
Cloud deployment models compared
| Dimension | Public cloud | Private cloud | Hybrid | Multicloud |
|---|---|---|---|---|
| Who uses the infrastructure | Many tenants on a provider's shared infra | One organization, dedicated | Mix of private/on-prem and public | Two or more public providers |
| Spending model | OpEx, pay-as-you-go | Mostly CapEx (you own it) | Blend of CapEx and OpEx | OpEx across providers |
| Primary reason to choose | Scale, agility, lowest TCO | Data residency, control, legacy | Bridge constrained + cloud-native workloads | Avoid lock-in, best-of-breed services |
| Main trade-off | Less low-level control | High fixed cost, limited elasticity | Network/integration complexity | Most integration and skills overhead |
Sharp facts the exam loves — give these one last read before exam day.
Cheat sheet
Sharp facts the exam loves — scan these before test day.
- Scalability means capacity can grow; elasticity means it grows and shrinks automatically
Scalability is the ability to add or remove capacity to meet demand, while elasticity is doing that scaling automatically and quickly in both directions as load rises and falls. Every elastic system is scalable, but a system can be scalable (you can add capacity) without being elastic (it does not adjust on its own). When a scenario describes capacity following traffic up and back down without anyone placing an order, the answer is elasticity, not just scalability.
Trap Answering 'scalability' when the scenario specifically describes capacity expanding and contracting automatically: that automatic, two-way behavior is elasticity.
7 questions test this
- A retail company is experiencing unpredictable traffic spikes during sales events. They want their containerized application on GKE to…
- An organization runs a web application that experiences unpredictable traffic spikes throughout the day. They want to automatically adjust…
- An organization wants to understand how cloud elasticity differs from traditional on-premises infrastructure when using Compute Engine.…
- A retail company needs their e-commerce platform to automatically adjust compute resources to match real-time demand during flash sales…
- An online retailer experiences unpredictable spikes in website traffic during promotional events. The organization wants its Compute Engine…
- An organization runs a web application on Compute Engine that experiences significant variations in user traffic throughout the day. They…
- An online retailer experiences large spikes in traffic during seasonal sales and very low traffic at other times. They want their computing…
- Reliability is staying up and recovering through failures, achieved by spreading across failure domains
Reliability is a system's ability to keep performing its function and to recover when something fails. In the cloud it is achieved by running across independent failure domains so that one fault does not take everything down, which is why workloads are spread across multiple zones. It is a property of the system, distinct from flexibility and agility, which are about the organization.
- Flexibility is freedom to change tech; agility is the business speed it unlocks
Flexibility is the freedom to choose, combine, and change technologies, regions, and providers without buying new hardware first. Agility is the business outcome that flexibility enables: teams ship and change products faster because they no longer wait weeks for servers to arrive. A question about launching features in days rather than quarters is describing agility; one about freely swapping technologies is describing flexibility.
- Cloud shifts spending from CapEx to OpEx
Capital expenditure (CapEx) is money spent up front on assets you own and depreciate, such as buying servers sized for peak demand; operating expenditure (OpEx) is money spent on ongoing consumption. In the cloud most resource costs are treated as OpEx: you pay as you consume, with no up-front hardware purchase. This CapEx-to-OpEx shift is the most-tested cost concept, and 'move from CapEx to OpEx' is usually the right answer when a scenario complains about large up-front hardware spend.
Trap Calling pay-as-you-go cloud spending CapEx: owned, up-front, depreciated hardware is CapEx; consumption-based cloud billing is OpEx.
4 questions test this
- A manufacturing company purchases physical servers every few years to run applications in its own data center. It is evaluating Compute…
- A manufacturing company has traditionally purchased physical servers every few years, requiring large upfront investments before any…
- An organization refuses to adopt cloud technology and continues to size its on-premises data center to handle its highest expected peak…
- A startup with very limited capital wants to launch its application without building or owning any data center hardware, paying only for…
- Pay-as-you-go is the default: pay only for what you consume
Pay-as-you-go is the baseline cloud pricing model, where you pay only for the resources you actually consume with no commitment and no up-front purchase. It is what makes the OpEx model work and is the right fit when usage is unpredictable or a team cannot commit to a term. Sustained-use and committed-use discounts then reduce the rate on top of this baseline.
7 questions test this
- A startup must launch a new application quickly but has limited capital and cannot afford to buy and provision its own servers. Which…
- A newly founded startup has limited funding and wants to launch its product quickly without purchasing or provisioning any physical…
- A development team deployed Cloud Functions to process messages from Pub/Sub. The function experiences variable traffic with periods of…
- A logistics company is comparing running workloads in its own data center against modernizing on Google Cloud. Which financial benefit best…
- An organization wants to understand how cloud elasticity differs from traditional on-premises infrastructure when using Compute Engine.…
- A manufacturing company has traditionally purchased physical servers every few years, requiring large upfront investments before any…
- A startup with very limited capital wants to launch its application without building or owning any data center hardware, paying only for…
- Committed-use discounts require a 1- or 3-year commitment; sustained-use discounts are automatic
A committed-use discount gives a lower price in exchange for committing to a one- or three-year term of usage, so it fits steady, predictable workloads you know will run for years. A sustained-use discount is applied automatically and grows the longer a resource runs within a month, requiring no commitment at all. The exam hinges on this contrast: 'we want a lower rate but cannot commit' points to sustained-use, while 'we will run this steadily for years and want the best rate' points to committed-use.
Trap Treating sustained-use discounts as something you sign up and commit to: they apply automatically with no commitment; only committed-use requires a 1- or 3-year term.
- Public cloud uses a provider's shared infrastructure for lowest TCO and most agility
Public cloud runs your workloads on a provider's shared, on-demand infrastructure such as Google Cloud, AWS, or Azure. It maximizes the OpEx model, elasticity, and agility and usually delivers the lowest TCO, with the trade-off of less low-level control over the underlying hardware. It is the default choice when scale, speed, and cost matter more than owning the hardware.
4 questions test this
- A startup must launch a new application quickly but has limited capital and cannot afford to buy and provision its own servers. Which…
- A newly founded startup has limited funding and wants to launch its product quickly without purchasing or provisioning any physical…
- An online media company expects highly unpredictable and occasionally viral traffic, and it does not want to own or operate any physical…
- A startup with very limited capital wants to launch its application without building or owning any data center hardware, paying only for…
- Private cloud is chosen for residency, control, or legacy, not for cost
A private cloud dedicates infrastructure to a single organization, often on-premises, and is chosen for data-residency rules, regulatory control, or legacy systems that cannot move. Its trade-off is mostly CapEx and limited elasticity because you still own and size the hardware. When a scenario cites a legal requirement that data stay in-country with no cloud option, private (or hybrid) is the conceptual answer.
- Hybrid cloud connects on-prem/private with public cloud so workloads span both
Hybrid cloud links a private or on-premises environment with a public cloud so workloads can run across both. It is the answer when part of the estate is constrained (say a regulated database that must stay private) while the rest moves to cloud-native services. The cost is the added network and integration complexity between the two environments.
4 questions test this
- An organization keeps its primary production systems running in its own on-premises data center but wants a cost-effective standby…
- A retailer already runs a critical legacy application in its own data center but wants to gradually extend specific workloads into a public…
- A manufacturer runs a latency-sensitive control system that must remain on the factory floor next to its production equipment, but it also…
- A financial services company wants to deploy new customer-facing applications in the cloud for scalability, but regulatory rules require…
- Multicloud uses two or more public providers, and needs a concrete reason
Multicloud means running on two or more public-cloud providers at once, justified by a concrete need such as avoiding vendor lock-in or using a best-of-breed service only one provider offers. It carries the most integration and skills overhead of any model, so it should be tied to a stated business reason rather than adopted by default. Note the distinction from hybrid: hybrid mixes private/on-prem with public, while multicloud mixes multiple public clouds.
Trap Choosing multicloud 'to be safer' with no stated reason: each extra provider multiplies integration and operational cost, so it must serve a concrete need like lock-in avoidance or a best-of-breed service.
4 questions test this
- An organization wants to avoid being locked into a single cloud vendor and intends to select best-in-class services from more than one…
- An organization wants to reduce its dependency on the availability of any single cloud provider by running comparable workloads across two…
- Two companies have merged, and each previously ran comparable production workloads on a different public cloud provider. Leadership decides…
- A large enterprise wants to run workloads across more than one public cloud provider to avoid dependence on a single vendor and to select…
- A region is an independent geographic area; a zone is a deployment area inside it
A Google Cloud region is an independent geographic area, and a zone is a deployment area for resources within a region. Each region contains three or more zones, each housed in a separate physical data center, and a zone should be treated as a single failure domain. You choose a region close to your users for lower latency and spread resources across zones so a single-zone failure does not take the workload down.
Trap Treating a zone as a whole region or assuming one zone gives high availability: a zone is a single failure domain, so resilience comes from spreading across multiple zones (or regions).
7 questions test this
- An organization needs to understand the relationship between Google Cloud regions and zones when planning their Compute Engine deployment.…
- A company is launching a new mobile application aimed at customers in a part of the world where it currently has no infrastructure. To…
- An organization needs to choose a Google Cloud region for deploying Compute Engine workloads. What is the primary benefit of selecting a…
- A company is explaining Google Cloud's infrastructure to its leadership team. How should the relationship between regions and zones be…
- An organization is selecting a Google Cloud region to deploy their Compute Engine workloads. They need to protect against hardware,…
- An organization is using Cloud Load Balancing with Compute Engine managed instance groups. A zone experiences an outage. What happens to…
- A retail company wants to reduce latency for customers accessing their e-commerce application. Their customer base is primarily located on…
- Multi-region services survive the loss of an entire region
Multi-region services such as Cloud Storage and Spanner replicate data across more than one region and are designed to keep working after losing a whole region. This is a stronger guarantee than multi-zone, which only protects against a single data-center failure within one region. Reach for a multi-region configuration when the requirement is to survive a full regional outage.
- Latency is delay; bandwidth is capacity: both describe network performance
Latency is the time it takes a packet to travel from source to destination, while bandwidth is how much data can move per second. They are independent: a link can have high bandwidth but high latency and still feel slow for interactive use. Placing resources in a region near users primarily reduces latency, which is why proximity matters for responsiveness.
Trap Assuming more bandwidth fixes a latency problem: bandwidth is throughput capacity, latency is delay, so a high-bandwidth but far-away region can still feel slow for interactive traffic.
5 questions test this
- A company is launching a new mobile application aimed at customers in a part of the world where it currently has no infrastructure. To…
- An organization needs to choose a Google Cloud region for deploying Compute Engine workloads. What is the primary benefit of selecting a…
- A global media company wants to reduce buffering and load times for users who stream video from many different countries. Which…
- An organization is describing its network needs and wants to identify the term for the amount of data that can be transmitted over a…
- A retail company wants to reduce latency for customers accessing their e-commerce application. Their customer base is primarily located on…
- Know the basic network terms: IP address, DNS, and ISP
An IP address is the numeric address that identifies a device on a network, DNS (Domain Name System) translates human-readable names like example.com into IP addresses, and an ISP (Internet Service Provider) is the company that connects you to the internet. The Cloud Digital Leader exam expects you to recognize these foundational terms rather than configure them.
- Google owns a global private fiber network so traffic largely avoids the public internet
Google operates a global, privately owned fiber-optic network (including subsea cables) that connects its regions, plus a worldwide set of edge points of presence (PoPs) in over 200 locations. Because Google owns this backbone end to end, user traffic enters Google's network close to the user and stays on Google's private fiber for most of its journey rather than crossing the open public internet, which lowers latency, raises throughput, and improves security. The business takeaway is that customers get global, low-latency reach without building any of this network themselves.
4 questions test this
- A global media company wants to reduce buffering and load times for users who stream video from many different countries. Which…
- A company wants user requests to enter Google's network as physically close to the user as possible so traffic can travel over Google's…
- A financial services firm is evaluating how Google Cloud transfers large volumes of data quickly and reliably between continents. Which…
- A company is concerned about inconsistent performance when routing traffic between its global offices over the public internet. How does…
- Pre-trained AI APIs add AI capabilities with no ML expertise or model training
Google Cloud's pre-trained APIs let you add AI to an application with a simple API call, requiring no data science team and no custom model training. Match the modality to the API: Vision API for images, Video Intelligence API for video, Speech-to-Text for audio transcription, and the Natural Language API for text analysis such as sentiment.
Trap Pre-trained APIs are for standard tasks out of the box; building a custom model on your own data without coding is AutoML on Vertex AI instead.
6 questions test this
- An organization wants to quickly add the ability to analyze customer sentiment from support emails without hiring machine learning…
- A global media company wants to add speech transcription capabilities to their video platform, supporting multiple languages. They want to…
- An organization wants to add image recognition capabilities to their mobile application to detect objects in photos. They have no data…
- A media company needs to automatically detect and tag content in videos for their digital archive. They have no machine learning expertise…
- A global company needs to transcribe customer support calls in multiple languages for quality analysis. They want to implement this quickly…
- An organization wants to add AI capabilities to their applications and needs to choose between pre-trained APIs and custom models. What is…
- BigQuery is a serverless data warehouse that decouples storage from compute
BigQuery is serverless: there are no clusters to provision, deploy, or manage, so analysts query large datasets with standard SQL and focus on data instead of infrastructure. Its architecture scales storage and compute independently for flexibility and cost control, supports streaming ingestion for near-real-time querying, and connects to BI tools like Looker, Tableau, and Power BI.
Trap Treating BigQuery as a provisioned warehouse that needs cluster sizing or capacity planning before you can query.
7 questions test this
- A retail company needs to make business decisions based on real-time customer transactions and market events. They want to respond…
- A company currently operates a traditional on-premises data warehouse with growing data volumes and escalating infrastructure costs. What…
- An organization wants to enable data analysts to focus on uncovering insights rather than managing database infrastructure. Which…
- An organization wants to analyze data from multiple sources without managing infrastructure or provisioning clusters. They need to enable…
- An organization wants to empower business users to analyze large datasets without requiring IT to manage infrastructure. Which BigQuery…
- A healthcare organization wants to integrate BigQuery data with business intelligence tools to enable self-service reporting across…
- An organization wants to analyze large datasets and make data-driven decisions without hiring specialized database administrators to manage…
- BigQuery ML and AutoML let teams build ML models without data science expertise
BigQuery ML lets SQL practitioners create and run machine learning models using familiar SQL, with no Python or specialized ML framework needed, democratizing predictive analytics for data analysts. AutoML on Vertex AI builds a custom model from your own training data code-free, automating data prep, model selection, and tuning.
Trap Reaching for custom-trained Vertex AI models and ML engineers when BigQuery ML or AutoML already meets the need.
4 questions test this
- An organization's leadership wants to enable business analysts who understand SQL to generate predictive insights without relying on data…
- A data analyst at a company wants to build machine learning models to predict customer behavior but has no experience with Python or…
- An organization needs to build custom machine learning models for their unique business data but lacks experienced data scientists. Which…
- An organization needs to build a custom image classification model to identify their specific product defects, but they lack ML expertise.…
- Cloud Billing reports visualize cost history, trends, and forecasts by project and service
Cloud Billing reports are the built-in dashboard for understanding spend: they break costs down by project, service, and SKU, show trends over time, and project likely future spend with a cost-trend forecast for budget planning. They also show savings from credits such as committed-use and sustained-use discounts, supporting departmental accountability and chargeback.
Trap Expecting Cloud Billing reports to alert you when spend crosses a threshold, which is the job of Budgets and alerts.
6 questions test this
- An organization wants to track its Google Cloud spending trends and understand which projects and services cost the most. They need to…
- A financial operations team wants to estimate their organization's future Google Cloud spending based on recent usage patterns so they can…
- An IT manager wants to drill into Google Cloud spend at the granular SKU level within a built-in dashboard to understand exactly which…
- A company's leadership wants a visual breakdown of cloud costs by project and by service so that each department can be held accountable…
- A finance team wants to understand how their organization's Google Cloud spending has trended over the past several months and identify…
- An organization wants to track its Google Cloud spending trends and understand how discounts are saving money on invoices. Which tool…
- The organization resource is the root of the hierarchy, and folders require it
The organization resource sits at the top of the Google Cloud resource hierarchy and gives central visibility and control over every project and resource a company owns. It is created via Google Workspace or Cloud Identity (and is provisioned automatically the first time such an account creates a project). Folders are an optional grouping layer between the organization and projects, so an organization resource is a prerequisite for using folders; IAM policies set on a folder are inherited by the projects inside it.
6 questions test this
- An organization wants to use folders to group projects by department and environment in Google Cloud. What is required before they can…
- A startup is new to Google Cloud and wants to understand the benefits of using the resource hierarchy with an organization node. Which…
- A company has multiple departments that each require their own set of Google Cloud resources with different access permissions. The company…
- A company is setting up its Google Cloud environment and wants centralized visibility and control over all cloud resources across multiple…
- A company has set up Cloud Identity for their domain. Which capability does this automatically provide for their Google Cloud environment?
- An organization wants to group projects into logical departments and environments such as development, testing, and production in Google…
- TCO comparisons must include the costs that get overlooked
A fair on-premises-versus-cloud TCO comparison counts all direct and indirect costs, not just hardware and software. Commonly overlooked items include the operational costs of running a data center (power, cooling, maintenance, and support) on the on-prem side, and ongoing, usage-based network egress charges on the cloud side, which replace what used to be a one-time cabling and network-hardware purchase.
Trap Reducing a TCO comparison to hardware and software prices while omitting power, cooling, staffing, and other indirect costs.