Scaling the application

Prior to the arrival of the public cloud, scaling applications to meet potential user demand required significant planning and upfront investments. A possibility of high load meant the need to provision spare hardware capacity to handle the maximum load well in advance of the event. Certain businesses, such as retail companies, which experience high load only during holidays, had to maintain large amounts of unused spare infrastructure during the entire year. Other businesses, such as utilities, which may experience high load due to unpredictable weather conditions, had to anticipate the maximum required capacity and run on over-provisioned infrastructure. Despite the very high cost of over-provisioned infrastructure, dynamic scaling was rarely implemented, as it required each organization to re-invent the wheel and required high collaboration between segregated...