Cloud computing is one phrase that has captured everyone's imagination in the 21st century. The debate on the topic of whether cloud computing is really a revolution or an evolution won't settle anytime soon. However, during the last couple of years, there have been multiple new start-ups based around this "new" form of technology, as well as some big players joining the league of service providers on the cloud.
Cloud computing is a way of delivering hosted services. However, it is more than just outsourcing, as it has to offer more flexibility, scalability, and automation. Another interesting aspect is self-service, wherein the consumer can request a VM on the fly, build an app and host it on the cloud, or request an infrastructure. Then the service gets provisioned in a transparent way. Of course, with the abilities, limitations, and possibilities with cloud computing, vendors have coined their definitions for the same, which may send wrong signals to the end consumer.
The National Institute of Standards and Technology (NIST) is the federal technology agency in USA that works with the industry to develop and apply technology, measurements, and standards. It published its definition of cloud computing, which has general acceptance among cloud adopters and IT gurus.
The link to download the documentation is http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdf
As per NIST, the cloud model is composed of five essential attributes, three service models, and four deployment models.
Every cloud model has these five attributes, regardless of the deployment or service models:
On-demand self-service: Consumers can request and get services provisioned without any intervention by IT teams or service providers.
Broad network access: There is support for most kinds of client platforms over the network.
Resource pooling: Providers pool their resources and allocate and remove the allocation to the consumers as per their requirement. This process is transparent to the consumer.
Rapid elasticity: Consumers notice that resources are available in abundance. These can be committed to them as and when required and released accordingly.
Measured service: Optimization and metering of resources for chargebacks.
Cloud computing services can be availed primarily as per the following service models:
Software as a Service (SaaS): This service model enables a consumer to use an application hosted by the service provider on the cloud rather than deploy it on their premises. Applications using this service model are messaging and collaboration apps, Office apps, finance apps, and a few others. Google Docs and Microsoft Office 365 are good examples of this model.
Platform as a Service (PaaS): This service model grants more flexibility to the consumer, and they can upload and deploy a custom app or database. They also get to control the configuration of the application-hosting environment. Cloud Foundry, one of the subsidiaries of VMware/EMC, is a PaaS provider.
Infrastructure as a Service (IaaS): In this model, the providers offer a subset of infrastructure that may consist of both virtual and, at times, physical machines, with complete control over the OS and installed apps and limited control over the storage and other networking components (host firewalls). There are a few contenders in this league, such as Amazon Web Services, Rackspace, and Microsoft Azure that provide both PaaS and IaaS service models.
A cloud environment setup is determined by factors such as cost, ownership, and location. So, there are different deployment models for different sets of requirements. The following are the deployment models for a cloud-based implementation:
Private: A private cloud is provisioned and dedicated to a single consumer. It can be managed by the organization on premise or off-premise, or it can also be run and managed by a service provider. Microsoft System Center suite of products assists the customer to set up an in-house private cloud with manageability over hypervisors from different vendors, namely Hyper-V, ESXi, and XenServer.
Public: A public cloud is provisioned and shared by many tenants and is managed by the service provider off-premise.
Community: A community cloud is a rare collaborative environment spanning across participants with a common objective. The participants are consumer organizations, and they put their resources under a common pool. This model is managed and maintained by one or more members of the community.
Hybrid: As the name indicates, this is a merger of two or more cloud models, and the usability is decided by the consumer. In principle, the cloud models are unique but connected by proprietary technology, and allow portability of relevant data between the models.
Cloud computing has generated a lot of excitement in recent years. However, it is still less mature than regular outsourcing. Nonetheless, the winning bid from cloud models lies in the concept of automation and self-service, giving the consumer the freedom of choice and manageability.
Moreover, we have seen some new acronyms being coined around service models in recent times, in addition to what has been stated earlier, such as XaaS and NaaS. For example, XaaS, or anything-as-a-service, makes the SPI (short for Software/Platform/Infrastructure) model converge as demanded and delivers it as a service. NaaS, or network-as-a-service, is based around network virtualization and allows provisioning of virtual network service to consumers. This is just an indication that cloud computing is changing and will be changing the face of IT in the times to come.