Understanding Cloud Characteristics - Deploying and Managing a Cloud Infrastructure - page 16

Resource Provisioning
Smart resource provisioning would mean determining the required compute, storage, net-
working, and other resources to spin up an application. This calculation would be based
either on a set configuration that specifies how much resources are needed or on a set pattern.
Installation and Configuration
Automated installation scripts or tools would have to be executed once cloud resources are
provisioned. Automated installation would mean installing and configuring all the required
third-party software, libraries, and tools the application needs. A common practice is to
write the installation script once and then package it inside the application so that once you
allocate a cloud instance or a set of instances, you simply deploy the app and it installs itself.
Integration
Applications, especially enterprise applications, seldom exist in isolation. Your CRM is
connected with your ERP, which is connected with your accounting software. In this
interconnected world of enterprise and consumer applications, you will have to make sure
that when you deploy a new application, it automatically takes care of all the integration
complexities that arise from it. Usually the applications are implemented in a modular way
such that the interconnected applications talk to each other through a generic protocol
such as, for example, a REST API. This way, deploying a new application or updating an
existing application will not break down the whole application ecosystem and would just
impact a part of the complete deployment.
Object Storage Concepts
To understand the scale of persistent data stored on the cloud and added every day, consider
this statement from Eric Schmidt, CEO of Google: “Every two days now we create as much
information as we did from the dawn of civilization up until 2003.”
This amounts to roughly over 5 exabytes (yes, exa ) of data that humanity has generated.
Most of this data is in either private or public clouds somewhere.
The scale is unique in the sense that never in humanity did we have the need to store
such a humongous amount of data in a centralized place or data center. This scale of data
presents unique challenges for both storing and read/write operations on the data.
File-Based Data Storage
Traditional file storage systems were created with human readability in mind. The hierar-
chical layout of the various legacy file systems was designed for a limited scale of storage.
There were several advantages of going this route:
Layered Organization of Data Files It's easier to have data (text, video, audio, structured,
and unstructured) files organized into layers with a hierarchical structure.