Understanding Elasticity, Capacity, and Service Model Maintenance

Introduction

The cloud computing model has the advantage of supporting highly flexible and dynamic resource utilization. Users of the cloud can consume as many or as few resources as they need, and they can adjust their consumption as their needs change. Cloud platforms do not automatically optimize resource allocation. The amount of resources cloud workloads require at any given time depends mainly on the user.

Capacity and storage

Capacity is defined as the workload capability, what a system can do. Now, the capacity becomes very important for the cloud because if we are dealing with our local data centers or network operation centers, we can go in there. We manage that space, plug in more server racks any time we want to, and so forth to make sure we can meet the capacity.

However, when discussing the cloud, we need to make sure our service providers can plug in more servers if they need to, or they have already got the servers to expand our capabilities as required. So the ability to grow is vital, and meeting our current demands is critical. Capacity is what can the system do? Moreover, it begins with how much space or storage I require to store our data?

Speed and concurrency

The second factor with storage and capacity is the speed of reading and writing operations, i.e., IOPS (Input Output Operations per Second). Thus, how many of these operations can we do in a second? Furthermore, we need to think about concurrency. It is not just about one process. It is about multiple processes needing to access the storage simultaneously. So that is storage capacity. Then we have the networking capacity, and here we are dealing with the speed of transfer and the number of users, so we need to make sure that each user can transfer data at the rate they require.

Understanding capacity with an example

If we have 20 users using cloud service and each of them has to wait every time they try to access a document for 10 seconds before that document is available to them. Then that is 200 seconds for all of those different users spread across ‘t’ amount of time. Now, imagine we access that document one hundred ...