Information Technology Reference
In-Depth Information
center and between data centers are interconnected through virtualized network switches.
This consolidation results in putting a complete infrastructure into a single resource pool.
Scalable
The promise of the cloud is infinite scalability. This is powered by complete data center
virtualization with hypervisors abstracting every underlying hardware component. New
tenants can be quickly added. With a public cloud, resources can always be scaled up to
accommodate usage spikes and quickly scaled down when the usage normalizes again;
tenants are billed for the exact resources they utilize.
Available
Because the complete infrastructure is treated as a single resource pool, the cloud imple-
ments true High Availability (HA). Hypervisors not only abstract the underlying data cen-
ter resources, they also abstract the physical location, enabling availability regardless of the
location from which a user connects. Physical agility can be controlled/limited in private
cloud implementations where the enterprise has control over the physical servers.
Portable
A hypervisor offers abstraction of the OS from the server on which it runs. This would
mean that you can take your VM from one server within a data center to another server
within the same data center or to a server in a different data center by the same vendor.
Network and Application Isolation
Before hypervisors, building networked applications required choosing from various protocols
and frameworks that often could not talk to each other. The Representational State Transfer
(REST) standard emerged out of the need to standardize the implementation of networked
applications. Because hypervisors virtualize the underlying networking components, applica-
tions need not be aware of the specific networking hardware or even physical addresses to
implement and deliver networked applications. This network hardware isolation also enables
smooth scaling and migration of applications across servers.
Figure 2.2 provides a quick example of how an application utilizes a backend RESTful
API running on Amazon's cloud without being aware of the underlying network hardware.
In the figure, the sample application interacts with the backend REST API, which makes
use of the underlying resources without being aware of the underlying network. All that the
application may need to be aware of is the TCP/IP address of the instances it's running on.
An application may also be deployed across data centers spread in different physical
locations without the need to modify any communication logic between different VMs
running the application.
Search WWH ::




Custom Search