[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"cheat-sheet---en":3,"domain-info---en":3,"topic-info----en":3,"next-cloud-computing-fundamentals-cloud-concepts-and-economics-cloud-economics-capex-vs-opex-en":4,"prev-cloud-computing-fundamentals-cloud-concepts-and-economics-cloud-economics-capex-vs-opex-en":260,"lesson-cloud-computing-fundamentals-cloud-concepts-and-economics-cloud-economics-capex-vs-opex-en":465},null,{"locked":5,"reason":3,"meta":6,"item":16},false,{"title":7,"description":8,"isFree":5,"estimatedMinutes":9,"difficulty":10,"learningObjectives":11},"Cloud Pricing Models","On-demand, reserved and committed-use, and spot pricing across AWS, Azure, and Google Cloud: the discount each one trades for a commitment, and which workload actually fits each one.",18,"beginner",[12,13,14,15],"Explain why cloud providers discount pricing in exchange for a usage commitment, and what a workload gives up to get that discount","Compare on-demand, reserved and committed-use, and spot pricing by discount size, commitment length, and interruption risk","Match a workload's predictability and interruption tolerance to the pricing model that fits it","Calculate the monthly savings a committed-use discount produces over on-demand pricing for a given spend",{"id":17,"title":7,"body":18,"description":8,"difficulty":10,"estimatedMinutes":9,"extension":183,"infographics":184,"isFree":5,"learningObjectives":194,"meta":195,"navigation":196,"path":197,"quiz":198,"seo":257,"stem":258,"__hash__":259},"courses/courses/cloud-computing-fundamentals/en/domains/01-cloud-concepts-and-economics/03-cloud-economics/02-cloud-pricing-models.md",{"type":19,"value":20,"toc":171},"minimark",[21,26,30,34,37,41,44,47,50,53,57,60,63,66,70,149,154,158,161,164,168],[22,23,25],"h2",{"id":24},"paying-the-same-rate-whether-you-need-flexibility-or-not","Paying the same rate whether you need flexibility or not",[27,28,29],"p",{},"Run a database server 24 hours a day, every day, for 3 years straight, and you are still charged the exact same per-hour rate as a team that might shut its server down tomorrow. On-demand pricing is built for that second team, the one that needs the freedom to walk away at any moment. If a workload never walks away, that built-in flexibility is not free. The team running it every single hour is quietly paying for an option it never uses.",[22,31,33],{"id":32},"on-demand-pay-for-exactly-what-you-use-whenever-you-use-it","On-demand: pay for exactly what you use, whenever you use it",[27,35,36],{},"On-demand pricing has no upfront payment and no long-term commitment. AWS bills EC2 On-Demand Instances by the hour or the second, with a 60-second minimum, at a rate set by the provider and unaffected by how long someone has been a customer. This is the model built for the exact problem the first lesson in this domain opened with: guessing how much capacity a brand-new or unpredictable workload will need. Guess wrong on-demand, and the fix is a few clicks, not a hardware return.",[22,38,40],{"id":39},"reserved-and-committed-use-pricing-trading-flexibility-for-a-discount","Reserved and committed-use pricing: trading flexibility for a discount",[27,42,43],{},"Commit to steady usage for 1 or 3 years, and a provider prices that certainty into a lower rate. AWS calls this Reserved Instances or Savings Plans, discounting up to 72% off On-Demand pricing, with a 3-year term always discounting more than a 1-year one because it removes more uncertainty from the provider's own planning. Payment can be split 3 ways: All Upfront, Partial Upfront, or No Upfront, and generally the more you pay upfront, the deeper the discount.",[27,45,46],{},"Reserved Instances come in 2 offering classes, and the difference between them is worth knowing precisely. Standard Reserved Instances lock in the largest discount, but the instance family and Region are fixed for the term; they can be modified within limits but never exchanged for something different. Convertible Reserved Instances trade some of that discount away in exchange for the right to exchange the reservation later for a different instance family, useful for a team that expects its needs to shift before the term ends.",[27,48,49],{},"Other providers sell the same underlying idea under different names. Azure calls its version Reserved VM Instances, also up to 72% off pay-as-you-go pricing, and adds something specific to its own business: Azure Hybrid Benefit lets a company that already owns Windows Server or SQL Server licenses apply them in the cloud, stacking with a reservation for a combined discount of up to 85%. Google Cloud calls the equivalent Committed Use Discounts, up to 55% off standard machine types and up to 70% off memory-optimized ones, in exchange for the same 1- or 3-year commitment.",[27,51,52],{},"Google Cloud adds a second discount that AWS and Azure do not offer at all: the Sustained Use Discount, applied automatically, with no commitment of any kind. Run an eligible VM for more than 25% of a billing month, and Google Cloud starts discounting it on its own, with the discount growing as the VM's runtime approaches the full month, up to about 30% off. On AWS and Azure, a discount only exists if you commit to it in advance. On Google Cloud, simply running a workload long enough earns one automatically. That is a genuine structural difference between how the 3 providers price commitment, not just 3 names for the same mechanism.",[22,54,56],{"id":55},"spot-and-preemptible-the-deepest-discount-with-a-real-catch","Spot and preemptible: the deepest discount, with a real catch",[27,58,59],{},"Spot pricing sells unused capacity that a provider would rather discount heavily than leave idle. AWS EC2 Spot Instances discount up to 90% off On-Demand pricing, but AWS can reclaim that capacity with only a 2-minute warning whenever it needs it for an on-demand or committed customer instead. Azure Spot Virtual Machines discount up to 90% off pay-as-you-go rates and offer no reservation option at all. Google Cloud Spot VMs discount up to 91%, and can be preempted at any time.",[27,61,62],{},"This model fits fault-tolerant, interruption-tolerant work: batch data processing, CI/CD pipelines, rendering jobs, anything that checkpoints its progress often enough that an interruption costs a few minutes, not a failed job.",[27,64,65],{},"It is tempting to treat spot pricing as simply a cheaper version of on-demand. It is not the same product at a lower price. It is fundamentally uncommitted capacity that the provider can take back whenever a paying on-demand or committed customer needs it instead. Putting a production database, something that cannot tolerate a sudden shutdown, on spot pricing is one of the most common cost-optimization mistakes teams make, and a frequent trap in exam-style scenarios for exactly that reason.",[22,67,69],{"id":68},"the-3-models-side-by-side","The 3 models, side by side",[71,72,73,91],"table",{},[74,75,76],"thead",{},[77,78,79,82,85,88],"tr",{},[80,81],"th",{},[80,83,84],{},"On-Demand",[80,86,87],{},"Reserved / Committed Use",[80,89,90],{},"Spot / Preemptible",[92,93,94,109,122,135],"tbody",{},[77,95,96,100,103,106],{},[97,98,99],"td",{},"Discount vs on-demand rate",[97,101,102],{},"None (this is the baseline)",[97,104,105],{},"Up to 72% (AWS, Azure), up to 70% (Google Cloud)",[97,107,108],{},"Up to 90 to 91%",[77,110,111,114,117,120],{},[97,112,113],{},"Commitment",[97,115,116],{},"None",[97,118,119],{},"1 or 3 years",[97,121,116],{},[77,123,124,127,130,132],{},[97,125,126],{},"Can be interrupted",[97,128,129],{},"No",[97,131,129],{},[97,133,134],{},"Yes, with little to no notice",[77,136,137,140,143,146],{},[97,138,139],{},"Best fit",[97,141,142],{},"Unpredictable, short-lived, or brand-new workloads",[97,144,145],{},"Steady, predictable baseline capacity",[97,147,148],{},"Fault-tolerant, interruption-tolerant batch work",[150,151],"infographic",{"alt":152,"slug":153},"A spectrum showing On-Demand, Reserved or Committed Use, and Spot pricing arranged by discount size, with discount growing as commitment and flexibility trade off against each other.","cloud-pricing-models-spectrum",[22,155,157],{"id":156},"worked-example-what-a-commitment-is-actually-worth","Worked example: what a commitment is actually worth",[27,159,160],{},"Say a workload runs steadily enough that its on-demand bill is $1,000 a month. Commit that workload to a 3-year Savings Plan at the ceiling AWS advertises, up to 72% off, and the same workload costs about $280 a month instead, a savings of roughly $720 a month, or about $8,640 over a year, in exchange for giving up the ability to walk away without penalty.",[27,162,163],{},"That discount only pays off if the commitment actually gets used. A 3-year Reserved Instance sized for a project that gets cancelled 8 months in is now a fixed cost with 28 months left to run and nothing left to run it on, which is exactly why committing should follow measured, steady usage, not a guess about future growth, the same guessing problem that made traditional on-premises IT so expensive in the first place.",[22,165,167],{"id":166},"where-this-leaves-you","Where this leaves you",[27,169,170],{},"Match the commitment to how sure you are, and to how much an interruption would actually cost. A workload running every day for the next 3 years belongs on a committed rate. A workload that can lose an hour of progress and simply restart belongs on spot. Everything in between stays on-demand until there is enough usage history to commit with confidence. None of these hourly rates is the full story, though: the price on this page is not the same thing as what a workload actually costs a company to run, once the people, the tools, and everything built around it get counted too. That gap is exactly what total cost of ownership measures, and it is where this topic goes next.",{"title":172,"searchDepth":173,"depth":173,"links":174},"",3,[175,177,178,179,180,181,182],{"id":24,"depth":176,"text":25},2,{"id":32,"depth":176,"text":33},{"id":39,"depth":176,"text":40},{"id":55,"depth":176,"text":56},{"id":68,"depth":176,"text":69},{"id":156,"depth":176,"text":157},{"id":166,"depth":176,"text":167},"md",[185],{"slug":153,"concept":186,"style":187,"aspectRatio":188,"labels":189},"A horizontal spectrum diagram with 3 positions along one axis labeled discount size, increasing left to right. On-Demand sits at the left end with an icon showing full flexibility and no lock. Reserved / Committed Use sits in the middle with an icon showing a calendar spanning 1 to 3 years. Spot / Preemptible sits at the right end with an icon showing a capacity block that can be pulled away. Each position carries a short caption with its discount ceiling and its main tradeoff. A footer strip carries the takeaway that discount grows as flexibility shrinks.","diagram","16:9",[190,191,192,193],"On-Demand: no commitment, full flexibility, the baseline rate","Reserved / Committed Use: 1 to 3-year commitment, up to 72% off","Spot / Preemptible: no commitment, up to 90% off, capacity can be reclaimed anytime","Discount grows as flexibility shrinks",[12,13,14,15],{},true,"/courses/cloud-computing-fundamentals/en/domains/01-cloud-concepts-and-economics/03-cloud-economics/02-cloud-pricing-models",{"passingScore":199,"questions":200},70,[201,210,216,223,231,241,249],{"question":202,"type":203,"options":204,"correctAnswer":206,"explanation":209},"What does a cloud provider get in return for offering a steep discount on reserved or committed-use pricing?","single",[205,206,207,208],"A promise that the customer will only ever use spot capacity","A predictable, committed usage pattern it can plan its own capacity around","A one-time setup fee paid through a reseller","Exclusive rights to the customer's application data","Reserved and committed-use discounts exist because a provider can plan its own data center capacity more efficiently when it knows, in advance, how much a customer will use and for how long. The customer trades flexibility for that certainty, and the provider prices the discount accordingly.",{"question":211,"type":203,"options":212,"correctAnswer":214,"explanation":215},"Amazon EC2 Spot Instances let you reserve guaranteed capacity for a fixed term, the same way Reserved Instances do.",[213,214],"True","False","Spot Instances have no reservation and no guaranteed term. They use spare capacity that AWS can reclaim with as little as 2 minutes of notice when it needs that capacity for on-demand or committed customers instead, which is the opposite of a Reserved Instance's guaranteed term.",{"question":217,"type":203,"options":218,"correctAnswer":221,"explanation":222},"A team runs a nightly batch job that reprocesses log files, checkpoints its progress every few minutes, and can safely restart if interrupted. Which pricing model fits this workload best?",[84,219,220,221],"3-year Reserved Instance, paid All Upfront","Dedicated Host","Spot or preemptible capacity","A job that checkpoints frequently and tolerates restarts is exactly what spot and preemptible pricing is built for: the deepest discount, in exchange for the provider's right to reclaim the capacity at any time. Committing this workload to a 3-year reservation would lock in a fixed cost for a job that does not need the guarantee.",{"question":224,"type":203,"options":225,"correctAnswer":226,"explanation":230},"What is the key difference between a Standard and a Convertible Reserved Instance on AWS?",[226,227,228,229],"Standard offers the largest discount but cannot be exchanged for a different instance family; Convertible offers a smaller discount but can be exchanged","Standard can only be purchased with a No Upfront payment option","Convertible Reserved Instances always cost more than On-Demand pricing","Standard applies only to spot capacity, and Convertible applies only to on-demand capacity","Standard Reserved Instances lock in the largest discount but can only be modified, not exchanged for a different instance family. Convertible Reserved Instances trade some of that discount for the ability to exchange the reservation later if instance needs change, which is the boundary the exam expects you to know between the two.",{"question":232,"type":233,"options":234,"correctAnswers":239,"explanation":240},"Which of the following are true about Google Cloud's discount model, based on this lesson? (Select all that apply.)","multiple",[235,236,237,238],"Sustained Use Discounts apply automatically once a VM runs more than 25% of a billing month, with no upfront commitment","Committed Use Discounts require a 1- or 3-year commitment, similar to AWS Reserved Instances","Spot VMs on Google Cloud can never be preempted once they start running","Sustained Use and Committed Use discounts can both be stacked on top of Spot pricing for the same VM",[235,236],"Google Cloud is unusual in offering an automatic discount, the Sustained Use Discount, with no commitment at all, on top of a committed-use option similar to AWS and Azure. Spot VMs can be preempted at any time, and discount types cannot be combined: a VM running on Spot pricing does not also collect Sustained Use or Committed Use discounts.",{"question":242,"type":203,"options":243,"correctAnswer":246,"explanation":248},"A workload costs $1,000 a month at on-demand rates. The team commits to a 3-year Savings Plan advertised at up to 72% off On-Demand pricing. Roughly what will the monthly bill be after the discount?",[244,245,246,247],"$720","$500","$280","$72","A 72% discount off $1,000 leaves 28% of the original cost, or about $280 a month. That is roughly $720 a month in savings, or about $8,640 a year, in exchange for committing to that usage level for the full 3-year term.",{"question":250,"type":203,"options":251,"correctAnswer":253,"explanation":256},"A company signs a 3-year Reserved Instance commitment sized for a project that gets cancelled 8 months later. What does this scenario best illustrate?",[252,253,254,255],"Reserved pricing always costs more than On-Demand pricing in the end","A commitment only pays off if the usage it was sized for actually happens","Spot capacity should have been used instead of a Reserved Instance","The provider will automatically refund the unused months","A Reserved Instance is a bet that the usage will continue for the full term. When the project ends early, the remaining 28 months still have to be paid for, turning what looked like savings into a fixed cost with nothing left to run on it. This is exactly why committing should follow measured, steady usage rather than a guess about future growth.",{"title":7,"description":8},"courses/cloud-computing-fundamentals/en/domains/01-cloud-concepts-and-economics/03-cloud-economics/02-cloud-pricing-models","m0ezqHWY_aaTLqn2jhlWvCKkCV0ftmkZ_lsXBk53fMU",{"locked":5,"reason":3,"meta":261,"item":270},{"title":262,"description":263,"isFree":5,"estimatedMinutes":264,"difficulty":10,"learningObjectives":265},"Hybrid and Multi-Cloud","How organizations connect private infrastructure to the public cloud, spread workloads across more than one provider, and why hybrid cloud and multi-cloud solve different problems despite getting used interchangeably.",17,[266,267,268,269],"Define hybrid cloud using the NIST framework and explain how it connects 2 or more distinct infrastructures","Define multi-cloud and distinguish it from hybrid cloud based on integration and infrastructure type","Explain why vendor lock-in and compliance requirements drive organizations toward hybrid or multi-cloud strategies","Evaluate a scenario to determine whether it describes hybrid cloud, multi-cloud, both, or neither",{"id":271,"title":262,"body":272,"description":263,"difficulty":10,"estimatedMinutes":264,"extension":183,"infographics":405,"isFree":5,"learningObjectives":414,"meta":415,"navigation":196,"path":416,"quiz":417,"seo":462,"stem":463,"__hash__":464},"courses/courses/cloud-computing-fundamentals/en/domains/01-cloud-concepts-and-economics/02-cloud-deployment-models/02-hybrid-and-multi-cloud.md",{"type":19,"value":273,"toc":398},[274,278,281,284,287,291,294,297,300,304,307,310,313,317,321,324,384,387,390,393,395],[22,275,277],{"id":276},"two-problems-two-different-fixes","Two problems, two different fixes",[27,279,280],{},"A regional bank keeps its core ledger on a mainframe in its own data center, because moving 40 years of transaction history and a strict regulator's audit trail is not something you do casually. But the bank also wants its new mobile app to run on infrastructure that can absorb a holiday shopping spike without buying servers it uses 3 days a year. That is one problem: connecting infrastructure you cannot leave with infrastructure you want to add.",[27,282,283],{},"A different company, a fast-growing social app, has already committed to a multi-year, multi-billion dollar contract with one cloud provider and watches that single contract dominate its infrastructure budget. Losing negotiating leverage, and being one contract dispute away from a painful renegotiation, is a different problem: depending on a single vendor for everything.",[27,285,286],{},"The deployment models from the previous lesson do not have a name for either fix on their own. Hybrid cloud and multi-cloud are what organizations build when public, private, and community cloud each solve part of the problem, but no single one solves all of it.",[22,288,290],{"id":289},"hybrid-cloud-2-infrastructures-bound-together","Hybrid cloud: 2 infrastructures, bound together",[27,292,293],{},"NIST defines a hybrid cloud as \"a composition of 2 or more distinct cloud infrastructures (private, community, or public) that remain unique entities, but are bound together by standardized or proprietary technology that enables data and application portability.\" Read that carefully: the pieces stay separate. The bank's mainframe does not become part of AWS, and AWS does not become part of the bank's data center. What connects them is a layer, built specifically to move data and workloads between the two, that makes the boundary as invisible as possible to the people using it.",[27,295,296],{},"That connective layer is a real product category, and several vendors sell exactly this. AWS Outposts extends AWS hardware, APIs, and services into a company's own data center, so a team can run the same tools on-premises and in the public cloud without learning 2 different platforms. Microsoft's Azure Arc and Azure Stack HCI do the equivalent for Azure: managing on-premises servers, Kubernetes clusters, and databases as if they were Azure resources, wherever they physically sit. In both cases, the point is portability, moving an application or a dataset across the boundary without rebuilding it.",[27,298,299],{},"Back to the bank: the ledger stays on the mainframe it cannot move, and the mobile app runs in the public cloud where it can absorb a traffic spike overnight. Hybrid cloud is what lets both halves work as one system instead of 2 unrelated ones, sharing identity, monitoring, and in some cases data, across the line between them.",[22,301,303],{"id":302},"multi-cloud-more-than-one-public-provider","Multi-cloud: more than one public provider",[27,305,306],{},"Multi-cloud means something narrower than it sounds: using public cloud services from 2 or more separate providers, AWS and Google Cloud, for instance, rather than committing everything to one. Google Cloud's own definition keeps it simple: an organization uses cloud computing services from at least 2 public providers to run its applications.",[27,308,309],{},"The social app from the opening is a real pattern, not a hypothetical. Snap Inc, the company behind Snapchat, committed to spend at least $2 billion with Google Cloud over 5 years starting in 2017, a contract so large that Snap's own IPO filing named it as a major financial risk. Snap's response was to add a second, separate commitment: at least $1 billion to AWS by the end of 2021, for what its filing called \"redundant infrastructure support.\" Snap's infrastructure today runs as microservices spread across both providers, so no single vendor's pricing, outage, or contract renewal can dictate every term.",[27,311,312],{},"Vendor lock-in is the main reason organizations go multi-cloud, but it is not the only one. Some services genuinely are stronger on one provider, so a team might run its data warehouse on one cloud and its machine learning pipeline on another to use each provider's strongest offering. Data residency rules in some countries require certain data to stay within that country's borders, and the fastest provider region there is not always the one a company already uses everywhere else.",[150,314],{"alt":315,"slug":316},"A side-by-side diagram contrasting hybrid cloud, private infrastructure connected to one public cloud, with multi-cloud, multiple independent public cloud providers with no connection between them.","hybrid-and-multi-cloud-comparison",[22,318,320],{"id":319},"the-boundary-hybrid-or-multi-cloud-and-why-both-is-common","The boundary: hybrid or multi-cloud, and why \"both\" is common",[27,322,323],{},"Hybrid and multi-cloud answer different questions, and a scenario usually only tests one at a time.",[71,325,326,338],{},[74,327,328],{},[77,329,330,332,335],{},[80,331],{},[80,333,334],{},"Hybrid cloud",[80,336,337],{},"Multi-cloud",[92,339,340,351,362,373],{},[77,341,342,345,348],{},[97,343,344],{},"What's connected",[97,346,347],{},"A private or on-premises environment plus at least one public cloud",[97,349,350],{},"2 or more separate public cloud providers",[77,352,353,356,359],{},[97,354,355],{},"Integration",[97,357,358],{},"Bound together deliberately: shared APIs, identity, and data portability",[97,360,361],{},"Often independent; providers usually do not talk to each other",[77,363,364,367,370],{},[97,365,366],{},"Primary driver",[97,368,369],{},"Compliance, latency, or infrastructure you cannot retire",[97,371,372],{},"Avoiding vendor lock-in, using each provider's strongest service",[77,374,375,378,381],{},[97,376,377],{},"Real examples",[97,379,380],{},"AWS Outposts, Azure Arc, a bank's mainframe plus a cloud-hosted app",[97,382,383],{},"Snap Inc running microservices across AWS and Google Cloud",[27,385,386],{},"A company can be both at once, and increasingly is. Picture a hospital network again: patient records stay on private, on-premises hardware for compliance, the hybrid half, while its billing system runs on AWS and its telehealth video service runs on a different provider chosen for lower latency in certain regions, the multi-cloud half. Nothing about the definitions forces a company to pick a lane.",[27,388,389],{},"One trap worth naming directly: running workloads in 2 different regions of the same provider, 2 AWS regions, for example, is not multi-cloud. Multi-cloud specifically means separate providers. Running in multiple regions of one provider is a resilience technique within a single cloud, a different concept covered later in this course.",[27,391,392],{},"A second misconception: multi-cloud does not automatically mean an application survives a provider's outage. Spreading unrelated workloads across 2 providers, one team's app on AWS, a different team's app on Azure, reduces contract and pricing risk, but if a single application only runs on one of those providers, that application still goes down when that provider does. True cross-provider resilience means actively architecting one workload to run on more than one cloud at once, which is harder and rarer than multi-cloud adoption numbers suggest.",[22,394,167],{"id":166},[27,396,397],{},"Ask 2 questions, not one, before you label a setup. Does it include a private or on-premises piece connected to a public cloud? If yes, that is the hybrid half. Does it involve more than one separate public provider? If yes, that is the multi-cloud half. A setup can answer yes to both, one, or neither, and the label only matters because it tells you what problem the organization was actually solving: keeping something it cannot move, or refusing to depend on anyone it does not have to. The next topic leaves deployment models behind and turns to the question every one of these choices eventually runs into: what it actually costs.",{"title":172,"searchDepth":173,"depth":173,"links":399},[400,401,402,403,404],{"id":276,"depth":176,"text":277},{"id":289,"depth":176,"text":290},{"id":302,"depth":176,"text":303},{"id":319,"depth":176,"text":320},{"id":166,"depth":176,"text":167},[406],{"slug":316,"concept":407,"style":187,"aspectRatio":188,"labels":408},"A two-panel diagram side by side. The left panel, titled Hybrid Cloud, shows a private or on-premises server icon connected by a solid bidirectional arrow to a single public cloud icon, with a small label on the connecting arrow reading data and app portability. The right panel, titled Multi-Cloud, shows 3 separate public cloud provider icons, labeled generically rather than by brand, positioned side by side with no connecting arrows between them, each tied to a different, separate workload. A footer strip carries the one-sentence takeaway distinguishing the two.",[409,410,411,412,413],"Hybrid Cloud: private or on-premises infrastructure connected to a public cloud","The connection: standardized technology that moves data and applications across the boundary","Multi-Cloud: 2 or more separate public cloud providers, usually independent of each other","Each provider serves its own workload, with no automatic integration between them","Hybrid connects what has to stay in place; multi-cloud diversifies who provides the rest",[266,267,268,269],{},"/courses/cloud-computing-fundamentals/en/domains/01-cloud-concepts-and-economics/02-cloud-deployment-models/02-hybrid-and-multi-cloud",{"passingScore":199,"questions":418},[419,425,429,437,446,454],{"question":420,"type":203,"options":421,"correctAnswer":334,"explanation":424},"A regional bank keeps its 40-year-old transaction ledger on its own mainframe for regulatory reasons, but runs its new mobile app on AWS to absorb unpredictable traffic spikes, sharing identity and monitoring tools across both environments. What does this describe?",[337,422,334,423],"Community cloud","Public cloud only","Connecting a private, on-premises environment to a public cloud, and binding them together with shared tooling so data and workloads can move between them, is the definition of hybrid cloud. Multi-cloud specifically means multiple separate public providers, which is not what is described, since only one public cloud, AWS, appears here.",{"question":426,"type":203,"options":427,"correctAnswer":214,"explanation":428},"Running workloads in 2 different regions of the same cloud provider counts as multi-cloud.",[213,214],"Multi-cloud specifically means using 2 or more separate providers, such as AWS and Azure. Running in multiple regions of one provider is a resilience technique inside a single cloud, not multi-cloud, and exam-style scenarios use this exact wording as a trap.",{"question":430,"type":203,"options":431,"correctAnswer":433,"explanation":436},"Snap Inc added a second, separate cloud provider commitment after its Google Cloud contract became a major financial risk. What is the primary driver behind this kind of decision?",[432,433,434,435],"Meeting a data residency law","Avoiding dependence on a single vendor","Connecting on-premises hardware to the cloud","Reducing network latency for end users","Vendor lock-in, being financially and technically dependent on one provider's pricing and contract terms, is the most common driver behind a multi-cloud strategy. It is exactly what pushed Snap Inc to add a redundant AWS commitment alongside its existing Google Cloud spend. Data residency and latency are real reasons some organizations go multi-cloud too, but they were not the driver in Snap's case.",{"question":438,"type":233,"options":439,"correctAnswers":444,"explanation":445},"Which of these are true about multi-cloud, based on this lesson? (Select all that apply.)",[440,441,442,443],"It always means an application automatically survives if one provider has an outage","It means using public cloud services from 2 or more separate providers","A common driver is avoiding dependence on a single vendor's pricing and contracts","Providers in a multi-cloud setup are usually independent and do not automatically integrate with each other",[441,442,443],"Multi-cloud means spreading usage across 2 or more independent public providers, most often to avoid vendor lock-in, and those providers typically do not integrate with each other automatically. The one false statement is the automatic-resilience claim: an application still goes down with its provider unless it was specifically architected to run across more than one cloud at once, which most multi-cloud setups are not.",{"question":447,"type":203,"options":448,"correctAnswer":449,"explanation":453},"A hospital keeps patient records on private, on-premises hardware for compliance, runs its billing system on AWS, and runs its telehealth video service on a different provider chosen for lower regional latency. How would you best describe this hospital's overall cloud strategy?",[449,450,451,452],"Both hybrid and multi-cloud","Multi-cloud only","Hybrid only","Neither; this is traditional on-premises IT","The private records connected to a public cloud satisfy the hybrid half of the definition, and using 2 separate public providers for different services satisfies the multi-cloud half. Nothing in either definition prevents an organization from qualifying as both at once, which is increasingly the norm rather than the exception.",{"question":455,"type":203,"options":456,"correctAnswer":460,"explanation":461},"According to NIST's definition, what specifically has to be true for 2 or more cloud infrastructures to count as a hybrid cloud, rather than simply 2 unrelated systems a company happens to operate?",[457,458,459,460],"They must be owned by the same organization","They must both be public cloud infrastructures","They must run identical software versions","They must be bound together by technology that enables data and application portability","NIST's definition requires the pieces to remain distinct but be deliberately connected by standardized or proprietary technology that lets data and applications move between them, tools like AWS Outposts or Azure Arc. Simply operating a private data center and a separate AWS account with no integration between them is 2 unrelated systems, not a hybrid cloud.",{"title":262,"description":263},"courses/cloud-computing-fundamentals/en/domains/01-cloud-concepts-and-economics/02-cloud-deployment-models/02-hybrid-and-multi-cloud","vUcflAyt2nSABthrpJ9Q3_VHTvO-A_-mK8ENTkCt0IM",{"locked":5,"reason":3,"meta":466,"item":475},{"title":467,"description":468,"isFree":196,"estimatedMinutes":469,"difficulty":10,"learningObjectives":470},"CapEx vs OpEx","Why buying a server and renting one from a cloud provider hit a company's books in completely different ways, and why a monthly invoice is not automatically the same thing as OpEx.",15,[471,472,473,474],"Define capital expenditure (CapEx) and operating expenditure (OpEx) by asking who owns the underlying asset","Explain why the cloud shifted IT spending from CapEx toward OpEx, and what that shift does to cash flow and risk","Identify the misconception that billing frequency, rather than asset ownership, determines whether a cost is CapEx or OpEx","Calculate the CapEx and OpEx totals for the same workload using real dollar figures",{"id":476,"title":467,"body":477,"description":468,"difficulty":10,"estimatedMinutes":469,"extension":183,"infographics":552,"isFree":196,"learningObjectives":560,"meta":561,"navigation":196,"path":562,"quiz":563,"seo":610,"stem":611,"__hash__":612},"courses/courses/cloud-computing-fundamentals/en/domains/01-cloud-concepts-and-economics/03-cloud-economics/01-capex-vs-opex.md",{"type":19,"value":478,"toc":544},[479,483,486,490,493,496,500,503,506,510,513,516,519,523,527,530,533,536,539,541],[22,480,482],{"id":481},"from-a-locked-in-guess-to-a-monthly-choice","From a locked-in guess to a monthly choice",[27,484,485],{},"The previous topic left you at the edge of a question every deployment decision eventually runs into: what does this actually cost? The first lesson of this domain already put a name on half the answer, in passing. A company that buys $40,000 of server hardware books that purchase as capital expenditure, or CapEx, months before it knows how much of that capacity it will actually use. Move the same workload to a cloud provider instead, and the same team pays for it differently enough that accounting uses an entirely different word: operating expenditure, or OpEx. The difference is not just vocabulary. It changes how much cash a company ties up, how a purchase affects its tax filings, and how much risk it carries into a decision that is hard to reverse.",[22,487,489],{"id":488},"what-capex-means-and-what-it-costs-you-upfront","What CapEx means, and what it costs you upfront",[27,491,492],{},"CapEx is money spent to acquire a long-lived asset, a building, a vehicle, server hardware, that the company then owns and uses for years. Accounting will not let you deduct the full purchase the day you pay for it. Instead, you depreciate it: spreading the expense across the asset's useful life on the income statement, even though the cash left the bank account all at once. A $40,000 server with a 5-year useful life, depreciated on a straight-line basis, shows up as an $8,000 expense every year for 5 years, but the full $40,000 was gone from the company's cash the month it was purchased.",[27,494,495],{},"That gap between \"cash spent\" and \"expense recognized\" creates 2 real consequences. First, the cash is committed immediately, before the business knows whether it will grow into that capacity or outgrow it. Second, the purchase is close to final: reversing it means selling used hardware, usually for a fraction of what it cost, not simply cancelling a subscription.",[22,497,499],{"id":498},"what-opex-means-and-why-cloud-providers-are-built-to-sell-it","What OpEx means, and why cloud providers are built to sell it",[27,501,502],{},"OpEx is the cost of running the business day to day, recognized as an expense in the same period it is incurred, with no lasting asset left on the balance sheet afterward. Renting compute from a cloud provider is exactly this: you pay for an hour of usage, and that hour's cost hits the books the month you used it, not spread across 5 future years of depreciation. Nothing about the arrangement requires you to own hardware, because you do not. The provider does.",[27,504,505],{},"This is also why AWS's own Well-Architected Framework lists \"adopt a consumption model\" as a foundational cost optimization principle: pay only for the computing resources you consume, and scale up or down as the business actually needs, rather than sizing a purchase around a guess about the future. The framework gives a concrete illustration: a development environment used only 8 hours a day on weekdays and switched off the rest of the time can cut its compute cost by as much as 75%, simply by not paying for the 128 hours a week nobody is using it. A CapEx purchase cannot do that. Once you own the server, it costs the same whether it is busy or idle.",[22,507,509],{"id":508},"the-misconception-billing-frequency-is-not-the-test","The misconception: billing frequency is not the test",[27,511,512],{},"It is tempting to assume that if a cost shows up on a bill every month, it must be OpEx, and if you pay a lump sum upfront, it must be CapEx. That rule breaks the moment you look at how cloud providers actually price commitments.",[27,514,515],{},"AWS lets you pay for a 3-year Reserved Instance or Savings Plan entirely upfront, in a single lump sum that can run into tens of thousands of dollars paid on day one. That single large payment still does not turn the bill into CapEx, because the company still does not own a server. AWS does. Accounting treats it as a prepaid operating expense, amortized across the 3 years it covers, not as a depreciating asset on the buyer's balance sheet. Run the comparison the other way, and it holds just as well: a company that finances a physical server purchase with a bank loan and repays it in 36 equal monthly installments is still paying for CapEx. Spreading the payments out does not turn a loan into rent. The company still owns, and still depreciates, the machine at the end of it.",[27,517,518],{},"The actual test is simple to state and easy to forget under pressure: do you own a depreciating asset that sits on your balance sheet, or are you paying for someone else's asset as you consume it? That question, not the invoice schedule, is what a finance team is really asking when it labels a cost CapEx or OpEx.",[150,520],{"alt":521,"slug":522},"A side-by-side comparison showing CapEx as one large upfront payment for an owned, depreciating asset, next to OpEx as many small payments each fully expensed as incurred.","capex-vs-opex-split",[22,524,526],{"id":525},"worked-example-the-same-workload-2-different-books","Worked example: the same workload, 2 different books",[27,528,529],{},"Say a company needs a database server to handle steady, predictable traffic for the next 5 years. It can go 2 ways.",[27,531,532],{},"The CapEx path: buy a physical server for $40,000. That $40,000 leaves the bank in month one. The income statement shows $8,000 of depreciation a year for 5 years. The company also now owns the rack space and power draw that server needs, and will eventually pay to dispose of it when it retires.",[27,534,535],{},"The OpEx path: rent equivalent compute from a cloud provider for $700 a month, or $8,400 a year. Each month's bill is fully expensed the month it is incurred. No cash is committed beyond that month, and the arrangement can be resized or cancelled at any time without owning anything to sell off afterward.",[27,537,538],{},"Add up 5 years of each path, ignoring financing costs for simplicity: the CapEx path spends $40,000 in total cash. The OpEx path spends $8,400 a year for 5 years, or $42,000. The cloud path here actually costs slightly more in total dollars over the full 5 years. That is worth sitting with for a moment: OpEx is not a guarantee of a lower price tag. Its advantage is smaller, reversible commitments instead of one large, irreversible one, which matters most when you are not yet certain the capacity will get used for the full 5 years. Whether the extra flexibility is worth that gap, and how to measure the full picture properly, is exactly what the total cost of ownership lesson later in this topic works through.",[22,540,167],{"id":166},[27,542,543],{},"Ask one question about any cost: do you own the asset, or are you renting the outcome? That single distinction, not how the invoice arrives, is what separates CapEx from OpEx. Knowing that cloud spending is OpEx does not yet tell you which OpEx option to pick, though. Cloud providers sell the exact same capacity several different ways, on-demand, reserved, spot, each trading a different amount of discount for a different amount of commitment. That choice is what the next lesson breaks down.",{"title":172,"searchDepth":173,"depth":173,"links":545},[546,547,548,549,550,551],{"id":481,"depth":176,"text":482},{"id":488,"depth":176,"text":489},{"id":498,"depth":176,"text":499},{"id":508,"depth":176,"text":509},{"id":525,"depth":176,"text":526},{"id":166,"depth":176,"text":167},[553],{"slug":522,"concept":554,"style":555,"aspectRatio":188,"labels":556},"A two-panel comparison diagram, CapEx on the left and OpEx on the right. The left panel shows a single large payment arrow hitting a server icon, with a small depreciation bar chart underneath spreading that cost across 5 shrinking yearly slices, labeled as a balance-sheet asset. The right panel shows a repeating sequence of small equal payment arrows hitting a cloud icon, each one fully absorbed the month it happens, with no asset icon left behind. A footer strip carries the one-sentence takeaway about the real test: ownership, not billing frequency.","comparison",[557,558,559],"CapEx: one large payment, an owned asset that depreciates over years on the balance sheet","OpEx: many small payments, each one fully expensed the month it happens, nothing owned","The real test: do you own a depreciating asset, or are you paying for a service as you use it?",[471,472,473,474],{},"/courses/cloud-computing-fundamentals/en/domains/01-cloud-concepts-and-economics/03-cloud-economics/01-capex-vs-opex",{"passingScore":199,"questions":564},[565,573,577,585,594,602],{"question":566,"type":203,"options":567,"correctAnswer":570,"explanation":572},"Which question actually determines whether a cost counts as CapEx or OpEx?",[568,569,570,571],"How often the invoice arrives","Whether the total cost is over a fixed dollar threshold","Whether you own the underlying asset or are paying for it as a service","Whether the provider happens to be a cloud company","Accounting draws the line at ownership: CapEx buys an asset the company keeps and depreciates, while OpEx pays for something consumed and expensed in the same period. Invoice frequency and dollar amount are surface details that often correlate with this split, but neither one is the actual rule.",{"question":574,"type":203,"options":575,"correctAnswer":214,"explanation":576},"Paying for a 3-year AWS Savings Plan entirely upfront in a single lump sum converts that spending into CapEx.",[213,214],"Even a large lump-sum prepayment stays OpEx, because the company still does not own any hardware, AWS does. Accounting treats it as a prepaid operating expense amortized across the commitment term, not as a depreciating asset on the buyer's balance sheet.",{"question":578,"type":203,"options":579,"correctAnswer":580,"explanation":584},"A logistics company takes out a bank loan to buy 10 physical servers and repays the loan in 36 equal monthly installments. How does its accounting department most likely classify this spending?",[580,581,582,583],"CapEx, because the company owns a depreciating asset regardless of the payment schedule","OpEx, because it is billed monthly in equal installments","OpEx, because a loan is not considered a capital purchase","It depends only on the interest rate charged on the loan","The company owns the servers the moment it buys them, and that ownership, not the monthly loan repayment schedule, is what makes this CapEx. A monthly payment plan changes the cash flow timing, not the accounting classification of the underlying asset.",{"question":586,"type":233,"options":587,"correctAnswers":592,"explanation":593},"Which of these are genuine effects of the CapEx model, according to this lesson? (Select all that apply.)",[588,589,590,591],"Cash leaves the business in one large amount before usage is proven","The purchase is easy to reverse if the business's needs change","The expense is depreciated across the asset's useful life on the income statement","The company owns the underlying hardware",[588,590,591],"CapEx ties up cash immediately, spreads the expense across years of depreciation, and leaves the company holding an owned asset. The one false statement is reversibility: a CapEx purchase is hard to undo, since reversing it means selling used hardware for a fraction of what it cost.",{"question":595,"type":203,"options":596,"correctAnswer":600,"explanation":601},"Why did the shift from CapEx to OpEx appeal to fast-growing startups in particular?",[597,598,599,600],"OpEx is always the cheaper total price over the life of a workload","OpEx eliminates the need for any IT staff","OpEx removes the need to ever pay for compute capacity","OpEx avoids tying up cash in a large purchase before the business has proven how much capacity it actually needs","A startup's biggest constraint is usually cash, not compute. OpEx keeps that cash free for a business still figuring out its actual usage pattern; it does not automatically mean a lower total price, and it does not remove the need to pay for compute or staff altogether.",{"question":603,"type":203,"options":604,"correctAnswer":606,"explanation":609},"A company buys a $40,000 server with a 5-year useful life and depreciates it on a straight-line basis. How much does this purchase reduce the company's reported profit on the income statement in year 1?",[605,606,607,608],"$40,000","$8,000","$4,000","$0, because the cash already left the bank","Straight-line depreciation spreads the $40,000 cost across the server's 5-year life, so year 1's income statement shows an $8,000 depreciation expense, even though the full $40,000 in cash left the bank account the day the server was purchased. That gap, cash spent immediately but expense recognized gradually, is exactly what CapEx accounting produces.",{"title":467,"description":468},"courses/cloud-computing-fundamentals/en/domains/01-cloud-concepts-and-economics/03-cloud-economics/01-capex-vs-opex","YiFuBEPhlYzupFZohR7I90ZXGIzpxJXpCGNAq7BCzH4"]