Understanding Cloud Characteristics - Deploying and Managing a Cloud Infrastructure - page 14

considerations is CloudSwitch. It operates along with VMware virtualization layers and
enables cloud administrators or data center engineering teams to automate the switch
between the enterprise's private cloud and an already configured public cloud.
Load Balancing
Implementing cloud bursting would allow for smart load balancing for performance-sensitive
applications. Whenever the requests increase the available capacity, the additional load is
shifted on to the public cloud, and as the number of requests for an application decrease, the
transparent switching mechanism would automatically move the load back into the private
data center along with all the persistent data.
In addition to demand, there are other operational scenarios in which cloud burst-
ing would fit well, such as, for example, upgrading components within the data center or
upgrading the complete data center. In such a scenario, the complete load of the enterprise
would be routed to the public cloud while the private cloud's data center gets updated.
Generally, this would involve experiencing downtime if the right cloud augmentation or
bursting strategy is not implemented.
Security
A hybrid cloud deployment may not enable cloud bursting or switching to a public cloud
every application running on top of it. In cases where applications operate on business- or
customer-sensitive data elements, the apps would have to be kept inside the controlled
premises of the private cloud data center to ensure security of the sensitive data. This
important aspect needs to be considered when a cloud switching strategy is implemented
in a hybrid cloud deployment.
Rapid Deployment
In an ideal world, or maybe in the near future, the IT world would have zero-time deploy-
ment, which means that you can construct a basic use case application and provision cloud
resources for it within no time. Rapid deployment, not zero-time deployment, refers to this
ability of a cloud vendor to instantly construct a whole stack of technologies, frameworks,
backend data bases, and third-party tools and software along with the required compute,
storage, and networking resources and deploy this in a fully automated way.
Traditionally, deploying enterprise applications and software involved several phases
and validation tests and required the involvement of a team of engineers to overlook the
deployment of the application. A competitive business landscape and the need to keep on
innovating has stripped even the biggest of enterprises of the luxury to have deployments
of applications take several days. This is true for the majority of the use cases that these
applications target, but not all. Applications that are critical to an enterprise may have a
rapid deployment strategy built in but would have rigorous validation and usability tests
before the deployment is complete.