Appendix C: Service strategy and the cloud

At the time of writing, a major new trend has been emerging. Known by various names and taking various forms – including ‘the cloud’, ‘converging infrastructure’ and ‘hybrid delivery’ – this trend will play a vital role in forming the strategy of IT service providers in the future.

It is clear that this trend is significant, but it is still too early to write an authoritative set of best practice guidance about how to define, manage and deliver service in the cloud. Many approaches have focused on the technology needed in order to embark on a cloud strategy (e.g. virtualization), and not on the strategic implications of using these technologies. Instead, this publication includes a brief overview of the major themes and issues surrounding this complex area, drawing on the work of Ewald Comhaire, Lee Kedrie, Ed Mangiaratti and Peter Gilis (all cloud strategists within HP).

The basic principle of the cloud is that whatever IT service or utility a customer needs can be provided directly using the internet (or intranet) on a pay-per-use basis. Customers do not see, nor do they care, how the services are created and delivered. All they care about is that each service meets a specific need at a specified price – and that it is scalable and accessible whenever and wherever they need it.

Cloud services are elastic, in that they can expand and contract (seemingly intuitively) to meet variable customer needs. Customers can demand and customize these services almost at will, and mix and match different components, all through easy-to-use self-service interfaces.

C.1     CHARACTERISTICS OF CLOUD SERVICES

Cloud services share the following five characteristics:

  On demand Customers can access them whenever and wherever they need them through the internet (or intranet).

  Ubiquitous access Any client can access these services, even thin clients, because they use standard networking mechanisms and protocols.

  Resource pooling Cloud services are provided by pooled resources which, because they are based on internet access, can be located anywhere. Physical and virtual devices are dynamically assigned and reassigned according to customer demand. Although customers do not have control over the exact location of these devices, they are able to specify high-level requirements. For example, they can specify that customer data be located in a specific country to comply with data privacy legislation.

  Rapid elasticity Services can be provisioned to rapidly scale up or down to meet real-time customer demand.

  Measured services Since cloud services are pay-per-use, it is important that their usage be metered, and that a financial model exists to be able to bill for these services. Resource usage is monitored, controlled and reported.

From these characteristics it is clear that service portfolio management plays a very important role in defining the service models, services and service packages that the service provider provides.

The main feature of cloud services is that the services in the service portfolio can be selected, combined and used in a number of packages – in real time as demanded by the customer. A cloud portfolio consists of clearly defined building blocks, each of which is a service in its own right, but customers determine how these services will be constructed and used.

Demand forecasting is essentially a function of market research, although the emphasis is not so much on anticipating demand as being able to respond to it rapidly. Resources are deployed and redeployed continuously for a variety of workloads.

Another key difference identified through these characteristics is the role of the customer. In more traditional service models, customer requirements are gathered and worked into clearly defined, single-purpose services. In cloud models, the customer is able to define and use services as their requirements evolve. Cloud service providers are able to respond quickly to changes in demand for both utility and warranty.

The pay-per-use aspect of cloud services requires sophisticated and flexible financial management processes and models. Profitable services are more easily identified, and non-profitable services discontinued, based on a response to customer demand. In many traditional service funding models it is almost impossible to determine the cost of an individual service, let alone whether it is contributing to the profitability of the organization. Cloud services are therefore easily perceived as having value because customers can link their outcomes to the cost of the service immediately.

C.2     TYPES OF CLOUD SERVICE

Apart from individual services provided to consumers and customers, cloud providers are able to offer the following types of service to IT organizations, providing them with flexibly priced sourcing options and reducing the need for high levels of investment in unused capacity. These types of service include:

  Software as a Service (SaaS) This is where a service provider enables a customer to use software without having to purchase the actual software itself. The customer is therefore able to benefit from the utility of the software, while paying only for the level of utilization that they require. An example of this is a service desk that uses third-party owned and hosted incident management software to log and resolve incidents over the internet. The brand of software is often irrelevant, as long as it provides the functionality and reporting required by the customer. It may not have the full custom functionality that suits every aspect of the customer’s requirements, but it meets enough of their requirements to be acceptable for the price paid. In this way the service provider can reach a broader range of companies with standardized functionality.

  Platform as a Service (PaaS) In this type of service, the service provider sells an environment and tools which can be used by customers to develop their own applications and services. Although the customer does not control the infrastructure and they do not own the tools and programming languages, they have full control over the applications.

  Infrastructure as a Service (IaaS) Here the service provider sells the use of infrastructure resources, such as storage or processing power. Users can deploy applications on this infrastructure, and use the infrastructure as a basis for internal or external services to their customers.

There is a danger with these types of service, however. Some IT service providers have used these options as a means to abdicate the responsibility and risks of the IT service to the customer. For example, one IT organization decided that it could not understand the business of its internal customers. They knew that the customer needed servers and storage, and so decided to offer these as services. In this way, they reasoned, they did not have to be concerned with funding, service level agreements or capacity management. What happened, however, was that the customer began insisting on control of their own ‘cloud’ services and investment in technology reached unprecedented levels. In addition, much of the equipment and storage provided did not meet customer needs because of the varying requirements for speed, refresh rates and flexibility. Had the IT organization understood its customers better, it would have been able to properly gear each cloud to their specific requirements.

C.3     TYPES OF CLOUD

There is not just one type of cloud. Service providers will need to decide which type of cloud service provider they want to be. Options include:

  Private cloud These are clouds provided by a cloud service provider to a single organization over a network that is owned by that organization or its service provider. The cloud could be provided by an internal service provider (internal private cloud) or an outsourcer (external private cloud). This type of cloud has many of the benefits of cloud (on demand flexible services) without many of the disadvantages (risk and security).

  Community cloud The cloud infrastructure is owned and operated by or on behalf of a group of organizations that share common concerns or practices. An example is a group of universities that provide a similar range of services to their students and academic staff.

  Public cloud The cloud infrastructure is owned by a service provider who sells a standard range of services to a number of organizations or to the general public.

  Hybrid cloud This type of cloud is where two or more of the types of cloud infrastructure, generally owned by different organizations, are combined for the specific purpose of porting data, applications or services.

C.4     COMPONENTS OF A CLOUD ARCHITECTURE

In order for a cloud service to be successful, there are at least four components that need to be in place. These are:

  Service catalogue and portal It is critical that customers and users know exactly what services are available and can access them easily. They must also know what the cost of the service is, and any terms of use. The service desk plays an important role in informing customers of information that is not readily available through the portal, and assists in providing access. In addition, some form of SLA dashboard will help to articulate what level of service can be expected for the money paid.

  Service governance This defines, communicates and automates the policies, standards and principles applied to the service provider in providing cloud services. Governance is defined and enforced through the cloud architecture, financial management, and service portfolio management.

  Service delivery management The most important part of this area is the ability to monitor and report on the utilization of the services, since this is how the cloud services are funded. In addition, service operation processes, capacity and availability management, and security and business continuity management are critical for the success of this component of the cloud architecture.

  Infrastructure and service delivery Here the emphasis is on the services themselves. This component of the architecture includes the service delivery mechanisms and service content, the middleware, the virtualized infrastructure and the facilities in which all these need to be housed.

Service strategy is key in defining the standards, governance, models and content of these architectures.