Have you ever approached your IBM architects, asking for recommendations on what solution to use, only to get the standard answer: “It depends.” Well, they’re correct. When choosing a platform to deploy an application, your organization will benefit from a fit-for-purpose solution vs. a fit-all model and platform. Your choice will depend on several factors, which will be examined here.
The examples provided here are derived from actual use cases. Using fit-for-purpose solutions, we’ve increased business agility, delivering measurable benefits. However, this article doesn’t describe the process of fit-for-purpose platform selection; it focuses on what factors to consider. For a more thorough evaluation, IBM provides a Fit for Purpose workshop that can help you consider all the factors important to you and assess them against various target platforms.
Regardless of the size of your organization, you likely have common strategic goals such as reducing operating costs, increasing business agility in enterprise applications, optimizing resources utilization by consolidating servers, being able to do dynamic capacity expansion and virtualization, and improving time to market for new features and capabilities.
By choosing an appropriate fit-for-purpose platform, you can standardize your IT capability and provide your users with rapid access to dynamically scalable IT resources. A fit-for-purpose strategy lets you assign workloads to best-fit environments and helps you achieve lowest overall cost per workload and build a foundation for cloud computing.
Consider why we need to consider fit-for-purpose platform selection. To support rapid growth and speed to market, a right-sizing approach should be used to optimally leverage the ideal platform based on best-fit deployment options. The platform you choose should be selected on an application-by-application basis and can be done for existing applications that aren’t meeting Service Level Agreements (SLAs) or for new applications. This will help you optimize resource usage and save money while providing dynamic scalability, high availability, and reliability.
Platform Selection Considerations
Start by asking these questions to help you understand the requirements and application performance characteristics:
• What are your hardware, software, licensing, resource, and business requirements?
• What are the characteristics of the application being deployed?
• What data and transactions will it access and where do they reside?
• What is data size and can data access be co-located?
• Does it need to integrate with any other workloads?
• Do you need to be able to prioritize this work?
• How does this application behave?
• What qualities of service does it require?
• Do you require dynamic workload management and capacity expansion capabilities?
• What activity will it exhibit—number of transactions, number of users, etc.?
• How predictable or unpredictable are these activity rates?
• Can the number of tiers be reduced to remove or reduce network latency?
• Is there a need to reduce deployment times?
• Is virtualization needed to make more efficient use of hardware and reduce cost?
• What are SLAs for availability?
• What are your performance, scalability, security, and manageability requirements?
Since every organization has different priorities, skillsets, standards and procedures, you may want to prioritize this list based on its importance to your environment.
Now that we know what to consider when selecting a platform, we can examine some choices.
Most large shops have a mix of mainframe and various distributed servers (see Figure 1). Based on application behavior and requirements, you can determine what platform may be the best fit. For example, the best candidates for zEnterprise include: