Testing and Deployment: Quality Is King - Deploying and Managing a Cloud Infrastructure - page 295

Technical Benefits
Cloud architectures offer the possibility of designing applications that could utilize the
power of bulk processing systems for various needs. Here are a couple of benefits:
Architecting Real-Time Processing Pipelines Document processing pipelines can be created
to convert documents between formats. This can also empower text indexing for fast search
and data mining over millions of records.
Image processing and video encoding/decoding pipelines can be created to process
multimedia data.
Architecting Batch Processing Systems Batch processing systems can be architected to
perform efficient and effective log analysis of machine-generated data for performance moni-
toring and auditing of large-scale compute systems. In addition, log analysis can be used to
generate regular reports for administrative and managerial purposes.
Other uses include nightly builds of updated source code repositories, including the
newest patches and updates. Another is to automate unit testing and deployment testing
for different deployment configurations.
Business Benefits
Cloud architectures offer some clear business-level benefits:
Quick Provisioning Coupled with Zero Upfront Investment No costs for purchasing
machines, storage, networking, power hardware, and operations staff. Organizations can
save time because management would not need to approve expenditure before a proposal
for new facility can be underway as it would with traditional architectures. The facility
currently under use for other operations can be utilized for provisioning new services.
Just-in-Time Infrastructure Dynamic resource provisioning helps users to control the
compute capacity required. Organizations do not need to worry about scaling their infra-
structure all the time.
Per-Use Pricing Utility-based computing comes with utility-style pricing, which allows
billing the customers for only the infrastructure and services that have been used.
Efficient Resource Utilization With the option of dynamic provisioning of resources,
users can relinquish cloud resources more efficiently by commissioning those they need
and decommissioning the ones not required anymore.
Parallelization Dynamic provisioning of resources also enables the users to provision
instances and use fully distributed and scalable applications for data processing, thus
reducing the time it would normally take to accomplish a task on a small number of
machines with a nondistributed setup.