Around the Cloud…
Because of the versatility of cloud computing, there’s numerous adaptive changes we are able to make to deal with potential performance problems with the deployed system. Figure 6 shows a few of these changes noted in designated circles. The very first change which comes in your mind when thinking about a higher response here we are at an application system around the cloud would be to increase the sources, mainly virtual machines (Figure 6, change 1). Unlike the Starbucks system, space limitations are less prominent in cloud systems. Although these limitations could be enforced by hardware (an actual server cannot accommodate thousands of virtual machines), it doesn’t always arrived at that. So when software development will, the program could be gone to live in an open cloud, which often has much bigger capacity than private infrastructure and space isn’t an issue. However, the issue then becomes certainly one of cost reserving increasingly more virtual machines means more income that should be compensated towards the public cloud provider. This can be a consequence, which we’ll discuss within the next part in regards to the financial aspects of scaling. Price is really the motive for that second change (Figure 6, change 2) getting rid of an online machine from the cluster when it’s not fully utilized. When the workload could be shared through the remaining sources, the spare virtual machine can be taken off in order to save costs.
Figure 6. The modified software development system using the changes as designated circles.
Another possible change concerns the redistribution of demands to specialized groups (Figure 6, change 3). We think that the burden balancer from the software system already distributes the incoming demands to specialized groups based on their requirements for particular sources. However, this is accomplished in an exceedingly straightforward manner and also the balancer transmits the CPU intensive demands towards the CPU cluster, the memory intensive demands towards the memory cluster and so forth.