January 25, 2015 | Written by: Mahesh Jadhav
Share this post:
Today, every information technology (IT) business wants to reap the benefits of moving its workload to the cloud. But as we all know, not everything is suited for cloud migration. In order to assess whether it’s beneficial to move to a given workload to the cloud, workload analysis is vital. Two common questions faced by business and IT architects are:
• What principles should I consider when determining what workloads are best suited for cloud?
• What is the best way to measure the criteria with minimal effort and impact on business?
Assessing workloads and aligning them to an appropriate cloud type (private, hybrid or public) and a relevant delivery model (infrastructure as a service, platform as a service or software as a service) is a pivotal step in cloud migration strategy that can ultimately determine the success of that strategy. If a comprehensive workload assessment is not completed, then migrating a workload to cloud could prove to be a hassle that may end in migration project failure and unhappy customers.
Many enterprise-level IT architectures consist of a number of different technologies, layers that have been added on over time to increase functionality of systems. This results in large and complex systems that are difficult to manage and assess for cloud suitability. Therefore, an attentive and effective workload assessment outline is required so that you can simplify and streamline the analysis. Without the right methodology, it can be difficult to know where to start. The larger the workload, the more complex and important it becomes to order the tasks.
The five-step workload assessment method
Let’s examine the five main considerations that you should examine while performing a workload assessment for cloud suitability. This method can be used to classify workloads and determine potential candidates for moving to a cloud.
1. Business impact: You will need to assess the workload and gauge its impact on your business.
• Perform a workload classification and determine its requirements (for example, test, development, pre-production and production)
• Understand how data flows through the enterprise and how it is related to the core services and business process for documenting the business impact of moving the application to cloud.
• Determine whether it is a business-critical workload. You would not want to move mission critical workloads first in cloud environment.
• Figure out if the need for deployment is urgent. Rapidly deploying services in the market might give your business a competitive edge; so identifying such workloads and marking them as a candidate for moving to a cloud could prove beneficial.
• Also, you’ll need to understand the availability of other business opportunities associated with the cloud computing for given workload.
2. Application architecture assessment: You should evaluate application readiness for a cloud environment.
• Make sure to assess the impact of application migration to cloud.
• How difficult or how cost effective is it to refactor the application for a cloud environment? For instance, does the application have hard coding for hardware?
• Consider scaling issues: can the application be scaled up by adding instances or only by adding more resource to a lone instance?
• Figure out how moving to cloud will impact application licensing and be sure to map to an appropriate cloud service model.
• Create application profiles and group the applications accordingly to ease the suitability process.
• Evaluate the application stack and design to find the best suitability for cloud.
• Determine if any existing or proprietary applications could hinder a migration strategy.
3. Technical characteristics: You should consider common integrations and dependencies, integration and interoperability, and customization and support.
• Verify the integrations and dependencies of given workload. A system that uses a cloud platform should consist of loosely coupled participants. This means that the communication interface between the participants is generic so that new participants can be added or changed with minimal disruption.
• Participants that are closely coupled shouldn’t be separated over a cloud. They may, however, both be hosted on the same cloud platform.
• Check the existing workload interoperability factor. Consider this factor for mapping the workload to services in cloud.
• Determine if any customization in the given application would impact its candidature for cloud. Also check the support factor in terms of application support document availability, its technical diagrams availability and so forth.
• Check if the application or workload is already virtualized or is virtualization ready, and determine its ease of migration to cloud.
4. Non-functional requirements (NFRs): NFRs play a crucial role in cloud workload assessment.
• Understand customer security concerns and regulatory compliance requirements.
• Check for any given workload availability requirements.
• Consider the business’ performance expectations for all applications.
• Figure out any backup requirements and the available maintenance windows.
• Determine if the ability to access the application from anywhere provides an advantage.
• Ensure that the plan to migrate an application to cloud is location-independent.
5. Support and cost: Evaluating the support resources required for given workloads and the operational migration cost could help in assessing the benefit of moving a particular application to the cloud. It can also determine if a given application really needs to be migrated to cloud or whether there would be no cost benefit in doing so.
You should also contrast the cost of hosting the system locally with the costs of hosting it on a cloud, given the required performance of the system.
Each of these five factors should be considered when determining whether a particular enterprise application is candidate for moving to cloud. At the end of evaluation, you should have a better understanding of which applications are best suited to move to cloud.
In a subsequent post, I’ll provide a sample workload analysis hierarchy for cloud suitability and guide you through mapping the workload to an appropriate cloud deployment and service model.
Please feel free to leave a comment below if you have any questions or to share your own experiences.