Read a sample: 1. A Retailer Outgrows Its Server Room
Maya Chen, operations director at fictional retailer Northstar Outfitters, has a problem that looks like good news. The company’s online store is attracting more customers, and its product team has ideas for new promotions, delivery options, and mobile features. Yet every promising idea seems to lead back to the same question: Can the server room handle it?
Northstar runs its website and inventory application on equipment it owns. When traffic rises sharply, pages slow down, and employees worry that customers will abandon their orders. To prepare for future promotions, the technology team proposes buying more servers. Maya hesitates. Extra equipment could help during busy periods, but it might sit underused for much of the year.
The team also needs a separate environment for testing changes. Creating one with the current process means requesting equipment, waiting for delivery, installing it, and configuring it. By the time a test environment is ready, a business opportunity may have passed.
Northstar is considering Amazon Web Services (AWS), a cloud provider that offers computing resources over a network. Maya does not need to become an engineer to take part in the decision. She does, however, need to understand what would change for the business—and what would not.
With its current on-premises approach, Northstar operates infrastructure at a location it controls. The company buys servers and related equipment, provides space and power, and assigns people to maintain the environment. Owning the equipment gives Northstar direct control over many infrastructure decisions, but it also makes the company responsible for planning capacity and maintaining what it purchases.
Cloud computing offers a different way to obtain technology resources. Rather than acquiring every server in advance, Northstar can provision resources from a provider as needed and pay according to the applicable pricing arrangements. Depending on the service, those resources might include computing capacity, storage, databases, or networking capabilities.
Provisioning means making a resource available for use. For example, an authorized team member could request a virtual server through an AWS interface instead of waiting for a physical machine to arrive. A virtual server provides computing capacity through software running on underlying physical hardware.
The cloud is not simply “someone else’s server room.” That phrase captures part of the arrangement, but it misses an important business distinction: cloud resources can be requested, changed, and released through services designed for flexible use. AWS operates the underlying infrastructure, while customers select and configure the services they need.
Maya’s decision is not whether technology will have a cost or require management. Either approach does. Her decision is whether the cloud’s flexibility and service options are worth the migration effort, ongoing spending, and changes to how Northstar operates.
Northstar’s first concern is capacity: having enough computing resources to serve customers and support applications. Too little capacity can cause slow responses or prevent customers from completing purchases. Too much can leave the company paying for resources it rarely uses.
On premises, Northstar can add servers, but the process involves lead time and a financial commitment. It must decide how much capacity to buy before it knows exactly what demand will be. If the company prepares for an unusually busy promotion, some of that equipment may be idle afterward. If it prepares for an average day, it may struggle during the promotion.
Cloud computing changes the options. Northstar can start with a chosen level of resources and increase it when demand grows. It can also reduce certain resources when they are no longer needed. The practical benefit is closer alignment between capacity and demand—not a guarantee that the alignment will happen by itself.
Two related terms describe this flexibility: scalability and elasticity.
Scalability is the ability of a system to accommodate growth by adding resources or otherwise increasing capacity. Northstar might need more capacity as its customer base expands over several months. A scalable design gives the team a way to support that growth.
Elasticity is the ability to adjust resources in response to changing demand, including reducing them when demand falls. Northstar’s traffic might rise during a promotion and decline afterward. An elastic approach aims to add capacity for the busy period and release the excess when it is no longer useful.
The concepts overlap, but their business emphasis differs. Scalability asks whether Northstar can handle more work. Elasticity emphasizes matching capacity to changing demand over time, including scaling down.
For a hypothetical illustration, suppose Northstar’s application normally needs four equivalent units of computing capacity and a promotion requires eight. An elastic setup might add four units for the promotion and later return to four. The illustration describes a capacity pattern, not a promised AWS configuration or bill. Actual resource needs depend on the application, traffic, and chosen services.
Northstar must still design and configure its application to use additional capacity. Adding servers will not necessarily solve a slow database, a poorly designed checkout process, or another bottleneck. The team also needs rules for when to add and remove resources, as well as monitoring to determine whether those rules work. Elasticity is an available capability, not a substitute for engineering and operational decisions.
Maya’s second concern is speed. The product team wants to test a new checkout feature, but the existing equipment is busy supporting current operations. Ordering hardware for a short-lived experiment feels disproportionate.
Agility is the ability to move quickly in response to business needs. In a cloud context, that can include obtaining technology…