Critical Design Elements for the Cloud

You cannot ignore it—the hype about cloud—it is everywhere. The key question is no longer what is the cloud? but rather what can I do in the cloud? If you believe the hype, the answer to that question is clear— you can do almost anything in the cloud.

About this Author

Michael Tainter

Director, Forsythe IT Service Management Practice

read profile | recent posts

Information technology (IT) organizations figure out how they can adapt and mature their IT design capabilities, and how they can take advantage of the shift from the traditional to utility service delivery to meet the demands of the business. When designing for the cloud or for any other service strategy, considering certain elements can lead to better success.

The ability of the cloud to deliver the services necessary to meet business expectations rests with the IT organization. While all IT departments define and design technologies to deliver services, few organizations address some of the more challenging elements required to ensure operable services meet their internal and/or external customers’ requirements. Elements such as establishing service level expectations, transition planning, supporting services in the cloud, and monitoring and management in the cloud.

Establishing Service Level Expectations

Any discussion relating to services must start with utility (fit for purpose) and warranty (fit for use). The basis for a service strategy is defining what the business wants to do in terms of expectations and requirements. By the time you arrive at the design stage, you should have already determined how the business uses services, the demand for those services and whether there are options in place to deliver them—or if a new design needs to be developed.

It is premature to consider cloud, or any other delivery option, before the service requirements are completely understood. A more mature method of defining requirements focuses on what the services will enable the customer to do vs. how the services will be delivered.

Early in the design stage is the time to determine and agree on the levels of service that will be expected by the business. A good design brainstorming session will address how the services are used: that is, the utility of the services. Questions to ask include what does the customer want to do and where, when, and how do they plan to access the services? For example, do they want to collaborate online? Quote, bill or order goods? This discussion will flesh out the customer experience in terms of how they use the service and their corresponding expectations. Once documented, these expectations can be used to facilitate a design that meets the requirements.

Understanding the service utility is the foundation for understanding the business needs.

The service requirements also should relate to the warranty of the services. Differentiating between purpose (what it is expected) and use is a key concept in designing service level agreements (value can only be achieved when both utility and warranty requirements are determined). Utility means it is available when you need it. And the warranty requirements focus on when the service is available, whether it has enough capacity (now and in the future), continuity requirements and whether the service is secure. Establishing understanding and gaining agreement on these terms is the basis for developing achievable service levels that deliver value.

Of course, defining these requirements is an iterative process as you move through the design phase and determine whether the requirements can be met. It is a good practice to evaluate the requirements throughout the service lifecycle to ensure they are meeting expectations. Take advantage of processes such as service catalog and service level management in the service design phase of the information technology infrastructure library (ITIL) v3 service lifecycle to help you establish the process you will follow to document services. The service catalog and the service design package are good places to document the utility and warranty requirements. Once fully documented, you now have the foundation for design considerations, including whether or not the cloud is an option.

Transition Planning

A second consideration for success involves planning the process and procedures for transitioning the services to operations, especially if cloud computing is an option. Solid technology solutions are a key differentiator in the marketplace, and generally, IT organizations are masters of technology. However, the right technology design is not enough. Planning how to manage the changes, releases and the information related to the new services also is critical to a smooth transition, as well as foundational to enabling operational support. Think about these questions in your design:

What types of changes will be required?

In answering this question, think about the service management framework of services, process, people and tools. Take into account that you are designing a service and not just the infrastructure that supports it. Processes and people need to be aligned in a manner that leads to success. Technology-minded people are accustomed to having tools help them deliver services, but’ does everyone understand how the tools will be leveraged to deliver a service? 

Can changes be bundled into an application release?

Reducing the impact of changes can often be achieved by bundling changes to lessen the frequency of major changes. A release approach includes designing the essential training and communications for a smooth transition. Appropriate levels of testing will be required to ensure that the defined service level requirements are being met. If these considerations are not addressed in the design phase, there could be a delay in transitioning the services to operations.

What information will you need?

This concept rests solely in the ability to document the configuration of the services and maintain them in a suitable configuration management database. The quality, scheme and use of the information are critical to success. Designing how the information will be used, accessed, and managed will help resolve any issues along the way. Transition planning for the cloud is no different than for any other solution; all aspects of moving the services to production must be addressed to be successful. What does change is the manner in which the services are provided. Since the cloud is a utility option for delivering services, it is critical for the IT organization to think more holistically regarding the service. The concept that cloud architectures are more converged, or boxed, if you will, implies that service provisioning should be faster and more agile. This consideration alone requires integration and bridge building between the traditional technology silos.

Supporting Services in the Cloud

Many organizations excel at designing for the technical functionality of the cloud, but few emphasize designing for support. The old saying “pay me now or pay me later” applies here. Cost optimization begins with the design for support. Every aspect of operations should be included, such as how to handle incidents, problems (root cause analysis), access to services, service requests, and the ability to monitor/manage events. The more thoroughly you design for operations, the better success you will have. Important questions to ask are:

How will service disruptions be managed?

As issues are identified, ensure you have a process to efficiently and effectively identify, diagnose, assign, resolve and close the issues. Make sure the IT service desk has all the necessary information to support the users. Your effort in planning for operations is critical in ensuring the user experience is managed when issues occur.

How will determination of root causes be handled?

Both reactive and proactive efforts to determine underlying causes of incidents will lead to improving the service reliability. Effective problem management, coupled with the tools to identify and manage performance issues will help lead to success. Your design must include addressing the process, people and tools that need to be aligned to eliminate further disruptions.

How are service requests going to be handled?

In the day-to-day use of the service, user requests will need to be supported. This capability should be a design consideration to determine which standard requests can be provided to the user community in an automated, self-service manner. As the types of requests are identified, the design must include the method in which these services are going to be provided and supported. Ironing out the workflow for these standard requests will ensure the operations team is aware of the process for fulfillment. Remember that speed of service provisioning is a key goal in the cloud. Defining and designing your support for the services will benefit the customer experience and increase your speed of delivery.

Monitoring and Management in the Cloud

Critical to success in providing services in the cloud, is the ability to design how the services will be managed. The ability to monitor the service’s performance is essential. Too often, management and monitoring solutions are developed after the services are operational. However, high-performing organizations define the monitoring requirements as part of the overall design of the service. They integrate these requirements with the service level requirements/objectives they define in their service level agreements. This approach helps ensure that you identify what you need to monitor, what you can monitor, and any related gaps before the solution is operational.

As a result, if there is a need for additional tool investments, you will have the information needed to make the right tool selection in preparation for the operational service, and determine any areas that will need various monitoring techniques.

In choosing a monitoring / management approach, consider the following three areas:

Service Monitoring

Since the service is being delivered as a holistic utility, ensure the monitoring solution can provide measurements that show overall service availability. This capability will require a solid service composition structure, or service tree, to be visible and allow for drill-down capability to see actual performance of the service, applications and infrastructure. The ability to document the service components is necessary in order to know the complete structure of the service. By taking a service-minded approach, you will allow for communication of service performance in business terms.

Application Performance

Delivery of the service will require an effective application performance monitoring capability. One area where application performance monitoring can be hugely influential is when coupled with testing requirements. As tests are designed and developed, leverage application monitoring to help prove that appropriate performance is being achieved and to avoid operational performance surprises.

End-User Experience

Every monitoring solution should include a mechanism to determine what the end user sees. This capability will pay dividends by showing how the service is delivered. This capability is the last mile of the service that is delivered to the end user. Designing this capability ensures that you know what the end user experience is before the service becomes operational, allowing for effective preparation prior to first access.

While there are many design elements to consider when providing services in the cloud, these four areas are often an after-thought. Addressing them early on, as part of your design, will help ensure you establish the capability to manage your services in accordance with your customer’s requirements. So, take the time to consider these concepts as part of the design phase; they will help elevate your organization’s success as you look toward the cloud.

An earlier version of this article was published in CIOUpdate.com.