It's been ten years after NIST clarify the difference between scalability vs. elasticity. But cloud elasticity and cloud scalability are still considered equal. For this reason, is both terms seem to be used interchangeably. But the definition of cloud computing is not complete without understanding the clear connection between both these terms.
Both of these terms are essential aspects of cloud computing systems, but the functionality of both the words are not the same? They are quite different as they seem.

Scalability vs. Elasticity

According to the definition of cloud computing, as stated by NIST in 2011, elasticity is considered a fundamental characteristic of cloud computing. In contrast, scalability does not. Elasticity does not exist without scalability.
Let’s get in more depth to understand the clear difference between scalability and elasticity in context with cloud computing to understand the clear distinction.
The elasticity of any system can have scaling capability according to the demand. That means that system can be scaled up or down. It is a sort of ability that is quite feasible to increase or decrease resources, and it is required before any system is declared a complete cloud.
That means that any system can be scalable even if it is not elastic.
When we need to increase the capacity of our IT infrastructure in general, we need to prioritize the system's performance before the cost of the system. That leads us to scale up the system.
The system must be elastic as well as scalable to run a cloud. If the system is not adaptable but is scalable, it does not comply with the definition of cloud. Therefore our system needs to have this capability but not necessarily to make use of it.
Sometimes elasticity can be related to infrastructure artificially as well as scalability to applications. The definition of NIST refers to capabilities and not application or infrastructure. These capabilities are less critical, but the overall system's ability to adjust should meet the variable requirement quickly.
Unfortunately, demand drops and spike quickly until the system support team is competitive enough to additional backup services online. There could have been an outrageous of servers, resulting in the losses of customers.
Until the backup system is scaled, there would have been massive business losses in lack of rapid alternatives. Therefore, infrastructure requirements should be fulfilled in no time using elasticity or scalability terms in no time than estimated for recovery of service shutdown.
End-user does not care about the technologies being used. As long as the servers and applications are up, things will go smoothly. The time it runs down, the quantum of scalability and elasticity is determined. This is done to maximize the uptime. Services and applications can be incredibly elastic, and an absolute need for additional hardware, real or virtual.

Moving From song cloud scaling to cloud elasticity

The reality of production load results in extreme scaling to compensate for the requirement. It is like joining a company that moves the application to the cloud. In this case, the system support team will do some work to make all the applications available, even during downtime. This process includes breaking the application into containerized services of the micro-level. It is required to optimize the performance to meet the estimated recovery time.
Implementation of a robust monitoring system to provide feedback related to the parameters of utilization and application performance.

Scaling Out and Scaling Up

Scaling out and scaling up means increasing the resources of a system like CPU capacity. Contrary to this, scaling down or scaling in is to reduce required resources or shrink down. It is also known as vertical scaling. This technique lets a single resource perform by increasing or decreasing its capacity.
For example, it could be the number of CPU cores, virtual or real. The scaling out or scaling up concept, also known as horizontal scaling, is a technique of cloning resources to meet the variable requirement.

Cloud Elasticity to the Rescue

Many cloud service providers have an automation tool that can be used for quick scaling response to meet cloud elasticity than cloud scaling. Because this automatic process is an appropriate response to load change, it is fast, which helps to eliminate the outrageous and idle time of servers.

When is Cloud Elasticity Required?

For a cloud to be a real cloud, rapid elasticity is required instead of just elasticity. It does not mean that your system should be 100% elastic. As long as it can be flexible, it's always an accurate cloud system.

It is critical for your system to have elasticity ready if you are running retail services like Christmas, Black Friday, Cyber Monday, or Valentine's day. Opposite to this, if your business is selling software or a small company with predefined growth throughout the year, it should not worry about the elasticity. Furthermore, you need to decide if you need to run your system on the cloud or not.
Still, if it is a cloud system, it should process the essential characteristics as per the definition of NIST to adhere to the primary cloud computing services. There is no certainty in the on-demand requirements, which makes elasticity very necessary for the cloud. If your system was down because of insufficient resources, then you may lose customers, so having elasticity on your cloud system is essential for your business to sustain.

Conclusion of Cloud Elasticity in Cloud Scalability

Now it is clear that the ability of a system to scale down or scale up is fundamental, but it is entirely different from its capability to respond quickly. The ability to scale up is not as efficient as reacting swiftly to a downtime or service shutdown.
The ability to scale up and scale down is related to how your system responds to the changing requirements. Elastically in the context of cloud computing, it is required that the scaling of the system is quick, and it means the variable demands that the system exhibit.
There should not a need for manual action if a system is a true cloud. The response system should be completely computerized to respond to changing demands.