The Importance of Elasticity in the Cloud
Benefits that cloud computing can bring to companies are many, and they can push businesses forward both from the technology and business perspective. Some of the main value propositions of the cloud are low startup deployment costs, pay-for-use model, scalability, and imediate time to market.
When talking about the advantages that cloud offers, elasticity is the one quality of cloud computing that has been increasingly driving cloud adoption among organizations. NIST defines elasticity in the following way: “Capabilities can be rapidly and elastically provisioned, in some cases automatically, to quickly scale out and rapidly released to quickly scale in. To the consumer, the capabilities available for provisioning often appear to be unlimited and can be purchased in any quantity at any time.”
The fact that IT resources can be added or removed rapidly provides companies with the ability to substantially improve operational efficiency. Cloud elasticity mirrors best in the fact that changes to the IT environments can be made in minutes which results in quicker time to market and more affordable IT infrastructures. With the cloud, companies don’t have to go through time-consuming and expensive process of buying additional hardware and software.
The elastic nature of the cloud allows for scaling of applications to handle peak loads and enables companies to avoid the downfalls involved with the fixed provisioning. Elasticity removes overprovisioning and underprovisioning of IT resources from the equation offering better cost optimization. The capability to scale up compute resources on demand is highly cost-effective as it requires additional operational expense vs capital expenditure which would traditionally be the case.
To handle peak loads, companies turn to over-provisioning anticipating more resources than they typically need which also leads to servers being idle during the downtime. Provisioning of these underutilized servers is a huge waste of money. Cloud computing, on the other hand, allows for scaling the resources down when the usage is low to better meet demand.
In the case of underprovisioning, companies anticipate fewer resources than they typically need, resulting in inability to meet the demand, poor performance due to overloaded infrastructure, and consequently lead to financial loss for a company.
Speaking at the recent Cloud Connect conference in Silicon Valley, California, Todd Papaioannou, VP Cloud Architecture at Yahoo! said that traditional resource provisioning when dealing with expanding services is a problem for big companies that run sites such as Yahoo! that deal with huge amount of traffic on a daily basis. It takes 3 to 4 months for them to get the new hardware. The cloud, on the other hand, will expand on demand based on increasing capacity.
Yahoo! Cloud supports 600M+ users, 200PB of data, 100Bn events per day. Papaioannou said that currently Yahoo! can apply elasticity for Async type workloads or batch workloads, but the struggle is with online workloads major reason for that being spin up time. The only way for Yahoo! to deal with spiking for online traffic at the moment is load shedding.
Watch a video of Todd Papaioannou’s keynote from Cloud Connect conference: