Best Practice Principles - Centralization

A key challenge associated with application delivery infrastructure is that it must support a distributed set of end-user devices, but be managed easily. Centralization means that configuration, management, and maintenance occur from a single point. This does not mean that all components of an application delivery infrastructure must exist in a single location. Instead, the key concept of centralization is that the components are logically centralized, but can be physically distributed.

Centralization allows organizations to realize numerous benefits, including:

• Simplicity of management.  By centralizing the IT infrastructure, less administration overhead is required to manage and maintain it.
• Improved data integrity.  Keeping data centralized allows greater control over content locking. With only a single, authoritative instance of a piece of data, there is less of a chance that data will be overwritten accidentally if multiple people work on it.
• Heightened security.  Defining the core set of the application delivery infrastructure and centralizing it, a standard set of security processes can be integrated into the architecture, thereby helping to reduce the risk of security threats.
• Reduced data duplication.  Centralization reduces the number of copies of sensitive files that must be managed.
• Faster change convergence.  Organizations must react to changing market conditions and user demands. Creating a centralized view of the entire ADI allows an organization to push changes across the infrastructure quickly and easily.
• Increased consistency.  Managing the infrastructure as disparate systems results in decreased consistency. As consistency decreases, end-users become aware of the infrastructure, and must start to navigate to avoid complexities. Being able to manage layers of the ADI as a single, central entity allows the organization to build consistency and simplify the infrastructure.
• Enabling efficiency and cost-reduction.  Having a decentralized infrastructure requires organizations to continuously recreate multiple aspects of the infrastructure. By centralizing the infrastructure, organizations can rely on the same infrastructure components to support multiple systems, helping to reduce the cost and the number of systems they need to maintain.

Whether or not to centralize and what to centralize are the questions that may plague organizations trying to plan their application delivery infrastructure. Characteristics of the environment will help dictate where centralization makes sense and where it does not. For example, network connectivity is an excellent gauge for determining whether or not centralization is feasible. If the connectivity to the central site is not large enough, fast enough, scalable enough, or reliable enough, centralization would be a hindrance to the business because, should the link fail, remote users would not be able to continue working. In these situations, there is a need to diverge from the centralization model, and support the distributed workforce through the remote and branch offices.

Even if part of the infrastructure is decentralized to support the business, there is still the need for partial centralization for the following reasons:

• From the user perspective, application delivery should appear to be a centralized system. Users should experience transparency through the infrastructure to their applications and data. Requiring users to remember different workflows based on their current scenario is not the most efficient use of their time.

Shared data must be managed centrally to avoid contention and integrity issues.  The infrastructure must be capable of tracking usage and maintaining the integrity of the data, by automatically and seamlessly dealing with contention issues.

• The decentralized environment should be managed centrally. As the layers of the ADI spread further apart, it becomes even more critical to bring the configuration and management of the infrastructure layers together into a simple, concise format. Having a centralized view and interface into the environment allows the organization to create granular policies that seamlessly provide the user with the most appropriate application delivery solution.

Centralizing the ADI does not mean that all components must be located in one place. Instead, centralizing means that end-users have a single point of access to applications and data. Therefore, applications should be made available from a single point, regardless of the user, device, network, or the physical location of the application or data.

Note:  In most circumstances, it is not recommended for users to be required to follow a different workflow based on their current scenario. This adds complexity and confusion into the environment.

Being able to view the application delivery infrastructure as an entity allows organizations to create efficiencies within the infrastructure. As the different layers of the ADI are brought together into a central decision matrix, autonomic responses can occur based on the current user scenario. The delivery of applications and optimization techniques should be dynamically set as a response to the system assessment of the scenario. These responses should be controlled through a centralized set of granular policies that are capable of aligning with the particular business needs.