Big Data Cloud Deployment Models
Two key cloud models are important in the discussion of big data — public clouds and private clouds. Cloud computing is a method of providing a set of shared computing resources that include applications, computing, storage, networking, development, and deployment platforms, as well as business processes. Cloud computing turns traditional siloed computing assets into shared pools of resources.
The two types of deployment models for cloud computing are public and private. These are offered for general purpose computing needs as opposed to specific types of cloud delivery models.
The public cloud
The public cloud is a set of hardware, networking, storage, services, applications, and interfaces owned and operated by a third party for use by other companies and individuals. These commercial providers create a highly scalable data center that hides the details of the underlying infrastructure from the consumer.
Public clouds are viable because they typically manage relatively repetitive or straightforward workloads. For example, electronic mail is a very simple application. Therefore, a cloud provider can optimize the environment so that it is best suited to support a large number of customers.
Likewise, public cloud providers offering storage or computing services optimize their computing hardware and software to support these specific types of workloads.
In contrast, the typical data center supports so many different applications and workloads that it cannot be easily optimized. A public cloud can be very effective when an organization is executing a complex data analysis project and needs extra computing cycles to handle the task.
In addition, companies may choose to store data in a public cloud where the cost per gigabyte is relatively inexpensive when compared to purchased storage. The overriding issues with public clouds for big data are the security requirements and the amount of latency that is acceptable.
All public clouds are not the same. Some public clouds are scalable managed services with a high level of security and a high level of service management. Other public clouds are less robust and less secure, but they are much less expensive to use.
The private cloud
A private cloud is a set of hardware, networking, storage, services, application, and interfaces owned and operated by an organization for the use of its employees, partners, and customers. A private cloud can be created and managed by a third party for the exclusive use of one enterprise.
The private cloud is a highly controlled environment not open for public consumption. Thus, the private cloud sits behind a firewall. The private cloud is highly automated with a focus on governance, security, and compliance. Automation replaces more manual processes of managing IT service to support customers. In this way, business rules and processes can be implemented inside software so that the environment becomes more predictable and manageable.
If organizations are managing a big data project that demands processing massive amounts of data, the private cloud might be the best choice in terms of latency and security.