Cloud computing is an emerging technology and IT service delivery area that should be evaluated and followed closely by SAP users over the next few years. Many aspects of cloud computing, such as Infrastructure as a Service (IaaS), can be used today for certain use cases. Some other areas require time to develop before large numbers of companies using SAP technology embrace them. For a detailed explanation of the four primary deployment models of cloud computing, see the sidebar, “Public, Private, Virtual Private, and Hybrid Clouds,” at the bottom of the article. The IaaS delivery model of cloud computing is increasingly a consideration for CIOs, especially given the inherent potential of new ways of architecting and managing computing resources. Whereas security, regulatory, and data protection issues must be addressed in the public cloud space, there are many cloud computing benefits that SAP customers can realize today. Common landscape challenges such as server sprawl and drawn-out approval and procurement cycles can be addressed by using public cloud services or building private and hybrid clouds. Temporary systems for training, proof of concepts, and other short-term initiatives can be deployed rapidly on public cloud infrastructures with limited capital expenditure and pay-as-you-go billing arrangements. IaaS-like technologies can also be used as the basis for an internal IT delivery strategy. Private and hybrid clouds offer similar functionality to public clouds with added security and control. In either case, system resources, staff, and expenses are used more efficiently, which adds flexibility and agility to organizations. Some benefits are realized strictly through the underlying virtualization technology (i.e., server consolidation, rapid provisioning, improved manageability, and low-cost high availability) while other features are made possible by cloud computing vendor offerings. These benefits include cost models that are considered operational as opposed to capital expenditures. Public clouds also offer unconstrained capacity and the ability to dynamically ramp up or ramp down computing resources as required. To help maximize benefits and minimize risk, SAP users who want to embrace cloud computing infrastructure services should consider using a staged approach, thereby allowing ample time for the staff to identify which applications and infrastructure can be effectively deployed and managed using cloud infrastructure technologies. It is important to determine when a public cloud infrastructure may be more appropriate versus private clouds or standard hardware options. Understanding evolving business requirements along with internal infrastructure costs, capabilities, and limitations is critical before considering cloud solutions. A logical first step would be experimenting with non-critical applications on a public cloud or with a controlled private cloud environment. Like any new technology, guidelines should be established to help ensure implementation and support are consistent with the requirements of the technology platform and the needs of the user organization.

SAP Landscape Challenges

SAP landscapes have grown larger and more complex through the years, increasing the need for infrastructure capacity, scalability, and adaptability for most clients. Through acquisitions by SAP and newly developed products, the number of available SAP applications has grown significantly. Typical landscapes have evolved from a single ERP instance to more complex environments with multiple integrated components such as SAP NetWeaver Business Warehouse, SAP Customer Relationship Management, and SAP NetWeaver Portal. Coupled with high-availability requirements and the need to maintain change management via multi-system configurations, this often creates a server sprawl that requires massive amounts of disk storage, additional support staff, and data center expansion including rack space, cooling, and power. The overlap of budget constraints and greater efficiency and performance needs tends to become a complex and confusing area. Organizations are under increasing pressure to control capital expenditures for infrastructure and related costs such as power. At the same time, they are looking for faster and more efficient ways to deploy and support increasing numbers of SAP systems. Temporary systems for training, proof of concepts, demonstrations, and other short-term initiatives are especially difficult to justify when capital expenditures or lease agreements must be approved, operational staff is overloaded, and data center capacity is limited. To help address these situations, many SAP users are optimizing their data centers using a number of solutions, and most have discovered that virtualization typically accounts for the majority of gains in efficiency and energy reduction.

Cloud Computing and IaaS

There are three primary delivery models of cloud computing: IaaS, Platform as a Service (PaaS), and Software as a Service (SaaS) (Figure 1). The model I am focusing on is IaaS as this provides compelling short- and long-term opportunities for organizations to optimize their SAP infrastructure while improving overall flexibility and agility for the business. In its simplest form, IaaS is the delivery of computer infrastructure (servers, storage, networking) as a service over a high-speed network, with automated provisioning and elastic capacity.   The fundamental component of IaaS is virtualization. IaaS virtualizes each layer of the computing infrastructure stack, including servers, storage, and networking. The layers can be allocated and configured dynamically without human intervention.

Public, Private, and Hybrid Cloud Examples

Amazon — Public Cloud

Amazon is one of today’s leaders in the public cloud marketplace and can serve as an example when describing public cloud IaaS capabilities. Amazon Web Services is a collection of remote computing services offered over the Internet by Amazon.com. Amazon’s IaaS offering is named Elastic Compute Cloud (EC2), which is a suite of services that allows a client to build a complete computing infrastructure (including servers, storage, and networking) via self service. There are no upfront costs and users pay for services as they use them. SAP administrators can build application system templates that are ready for rapid deployment to on-demand instances when required. The first step is building or leveraging an existing Amazon Machine Image (AMI) with the appropriate operating system such as Linux or Windows Server. SAP applications are installed and patched to complete the template, which may be stored on Amazon’s Simple Storage Service (S3). A single image can be deployed to multiple instances and each instance can be turned on in minutes rather than days or weeks. Instance charges vary by size and accrue on an hourly basis with additional fees for Internet data transfer to and from the instance. When the instance is turned off, the only recurring fees are for storage of AMI templates on S3.

VMware — Private Cloud

Organizations running SAP that wish to build private or hybrid cloud environments in their data center can leverage virtualization and cloud management solutions provided by vendors such as VMware. VMware is an SAP Global Partner and its virtualization platform is fully supported by SAP for Windows and Linux configurations including production systems. In its simplest form, a VMware-based private cloud consists of SAP instances running on virtual machines. Immediate benefits include server consolidation, rapid provisioning, and high availability. Additional software, automation, and testing is required to use advanced functions such as self-service provisioning with approval workflow, bursting, metered usage with chargebacks, and detailed usage reporting.

VMware + Public Cloud Services — Hybrid Cloud

Building on the previous examples, combining a private cloud based on VMware virtualization technology with public cloud services such as Amazon can be used to create a basic hybrid cloud. Combining VMware with Amazon is limited today due to server image incompatibilities and completely different management tools; however, there are public cloud vendors that offer full compatibility with VMware’s server format. This compatibility provides flexibility in deployment and moving systems between on premise and off premise architecture. Additional information on hybrid clouds can be found in the section “Building an SAP Hybrid Cloud.”

SAP Use Cases of Cloud Computing

In an SAP landscape, public cloud IaaS provides the ability to rapidly deploy new servers with storage, no upfront capital expenditure, and no long-term commitment. It can support projects that would typically require much longer lead times to procure equipment and acquire approval via traditional methods. Public cloud IaaS is suited for systems that require quick deployment with short-term durations. I believe SAP sandbox, test, training, and demonstration systems are suitable candidates for public cloud IaaS. SAP users can also use private and hybrid clouds to add flexibility and security to their SAP landscapes. Private clouds offer similar benefits to public clouds with more control over data, security, and quality of service. Since the organization owns the infrastructure, it has control over how the applications are deployed and accessed, and how the cloud is managed. Another tool available to SAP users is the Adaptive Computing Controller (ACC), which abstracts application services from specific servers and provides operational, observational, and management control of entire landscapes from a central point. Features include the ability to start, stop, and dynamically relocate services or entire servers to optimize use. SAP ACC 7.2 now supports hypervisor-based virtualization solutions by SAP associates so these features are available for both physical and virtual server architectures. The SAP ACC 7.2 media is delivered as part of SAP NetWeaver and components are installed as an add-on to a SAP Composition Environment 7.2 system. An SAP user can now combine private cloud configurations using virtualization along with SAP’s Adaptive Computing Controller to reduce operational costs and increase the flexibility of their landscape.

Detailed SAP Use Case

An SAP customer is considering an upgrade to the next SAP ERP release. It would like its functional team to evaluate the new functionality first to build a business case for the upgrade. The technical team would like to practice the upgrade procedure using a copy of the production system to identify potential problems before the actual upgrade. Finally, the company would like to deliver training to 2,000 users before the upgrade begins. Situation: No physical servers are available in-house. However, the company has been experimenting with virtualization in its data center and Amazon public Web services. An in-house VMware cluster with plenty of disk was configured to begin development of a hybrid cloud and several Amazon machine images were created on Amazon’s public cloud, including one for the new SAP release. The company must start the evaluation promptly to meet project approval deadlines. The following examples are used to depict possible uses of public and private clouds. The most effective approach depends on several factors such as length of time the systems are required, internal infrastructure capacity, cost comparison, and data sensitivity. Solution for functional team: An Amazon instance running a fresh installation of the new release can be deployed in less than an hour with the appropriate amount of disk space. Since this is located on the public cloud, global team members can access the system via the Internet using SAP GUI and evaluate the new release. If outgoing TCP/IP ports are blocked in accordance with internal security policies, a virtual private cloud can be used to resolve this connectivity issue. An alternate approach would be to deploy an in-house virtual machine on the VMware cluster. Solution for technical team: Since a copy of production will be used, the company has concerns regarding exposure of sensitive information and will not use Amazon public cloud services; however, they are able to deploy an in-house virtual machine on the VMware cluster without purchasing additional hardware. A copy of production with a new SAP SID is then ported to the new virtual machine via a homogeneous/heterogeneous copy or faster methods developed in-house. System resources such as virtual CPUs or memory can be increased easily to meet performance requirements. Leveraging in-house virtualization enables rapid deployment while ensuring more control and security. Solution for training requirement: No training servers exist so the company can deploy several Amazon instances to accommodate user training. Since there are limited capital expenditures and no upfront costs, systems can be brought online immediately and shut down when the training is complete. Costs are minimized using the pay-as-you-go method. No local data center space, power, or cooling is used and operational staff interaction is minimal or not required. If the training instances include sensitive data, the internal VMware cluster can be used to deploy training instances instead of the public cloud.

Challenges with SAP on Cloud Computing IaaS

A major benefit of cloud computing IaaS and the underlying virtualization technology is the ability to quickly and easily clone systems. With cloud technology, it is possible to clone an existing SAP system and its database with minimal effort in a very short period of time. This creates a dilemma in that the new system will have the same SAP SID and hostname of the source system which can lead to user confusion and network conflicts if not properly managed through system access controls. Currently, SAP only supports the homogeneous/heterogeneous system copy process to change the SID and hostname of an SAP system. This method is a manual and time consuming process, so it negates the benefits of fast cloning. Because SAP currently does not provide a method to easily change the SID and hostname of a system, customers must develop their own method, which may not be supported by SAP, rely on third-party solutions, or isolate systems via network configuration to fully leverage the benefits of cloud IaaS’s quick and easy cloning capabilities. SAP supports the use of private clouds for production systems (as long as the underlying technologies are certified and released by SAP and any automation mechanisms do not have a negative effect on SAP operation) but does not currently support the use of public cloud IaaS for production systems as per SAP Note 1380654 which states “…the operating of SAP systems in a public cloud has not been released by SAP.” Non-technical areas that require attention due to lack of established standards in the public cloud space include licensing models (SAP and third party), data privacy laws, data protection, and service level agreements. SAP has not yet released detailed statements regarding support of non-production systems in a public cloud. This should not discourage SAP customers from experimenting with public cloud services and leveraging these capabilities for non-production systems as described in the previous section.

Building an SAP Hybrid Cloud

Hybrid clouds provide flexibility for organizations running SAP by leveraging benefits of both private and public cloud models. The following steps provide a high-level approach to building a hybrid cloud: Step 1. Evaluate internal infrastructure costs, capabilities, and procedures. Understanding the current situation in detail is essential before considering alternative deployment and management strategies for your SAP landscape. Determine the people resources, cost, and time required to deploy, maintain, and decommission an SAP instance within the organization. Examine procurement and approval procedures along with data center operations associated with new and existing SAP architecture. This information serves as a baseline for comparison to public and hybrid cloud configurations. Step 2. Determine requirements and select an approach. Review business drivers and current challenges that a hybrid cloud can help overcome. Evaluate skill sets and availability of internal staff to implement and maintain cloud architecture. Consider working with a consultant who can provide advisory and/or people resources for the project. Compare cloud management software and cloud in a box vendor offering to your IT staff’s ability to build a hybrid cloud. Step 3. Implement and optimize a virtualization strategy for SAP systems. When building your own hybrid cloud, the next step is to standardize the environment and introduce virtualization. If a virtualization strategy is not in place, implement an SAP-supported virtualization stack and begin testing less critical systems. When possible, you should align this process with an overall green initiative or virtualization implementation for the entire organization that considers SAP-specific resource requirements. Complexity can be minimized by standardizing hardware, operating systems, and database platforms as much as possible. Once virtualization is in place, the SAP system must be optimized for the virtual environment. Implement configuration leading practices published by the virtualization vendor and SAP, then test applications to ensure system performance and availability meet service levels required by the business. Develop SAP-specific cloning procedures and automation techniques to help ensure rapid and consistent system provisioning. Keep track of the time it takes to provision SAP instances for comparison to traditional methods or vendor solutions. Build a template library of SAP applications coupled with a strategy that includes methods for storage and keeping templates inventoried, up to date, and backed up. Step 4. Evaluate and test public cloud capabilities with SAP applications. Some public cloud offerings can be evaluated without establishing a formal relationship with the vendor. A credit card is generally all that is required to get started with a pilot project. Determine SAP applications and system types that are candidates for use on public cloud and install these on public cloud architecture. Do not place sensitive data on public cloud architecture without first performing an appropriate risk analysis. Develop a scorecard to compare how vendor offerings stack up against core requirements. Evaluate each vendor’s toolset used for deploying and managing architecture. Test SAP operations and user workloads on different instance sizes to determine the most effective configuration that meets required performance levels at the lowest cost. Keep track of the time it takes to provision and decommission SAP instances. Compare public services to virtual private cloud services by developing and testing use cases for both. Read and thoroughly understand service level agreements, support options, and cost models before selecting a vendor.

Putting It All Together

After virtualizing the internal SAP environment and selecting a public cloud vendor, practice and master baseline functions for both environments such as provisioning, backup/recovery, performance tuning, and decommissioning SAP instances. The next step is finalizing appropriate use cases for each environment. For example, when sensitive data or additional control is required, internal infrastructure may be more appropriate than public cloud services. Other factors may include cost, capacity, performance, or project duration. If desired, these criteria will drive automated provisioning and approval workflows that can be developed in-house or installed as part of a vendor solution. One of the most challenging steps is bridging the private and public infrastructures to form an integrated and automated mechanism for system provisioning and management. Vendor solutions are starting to evolve in this area and should be evaluated and compared to organizational requirements and IT staff capabilities for developing an in-house solution. Optional functions include centralized management consoles, automated bursting, self-service provisioning with approval workflow, metered usage with chargebacks, and detailed reporting. A high-quality solution supports physical and virtual server workloads and allows instance provisioning, management, and migrations across private and public infrastructure.

Cloud Recommendations

Although few organizations using SAP systems are ready to move their mission-critical enterprise applications to public cloud providers, they are starting to embrace public cloud infrastructure for important use cases as part of the overall IT strategy. Key benefits of public clouds include reduction of internal operational workloads and increased flexibility in computing, staff, and capital use. The use of private clouds is also growing significantly, providing benefits of a public cloud with more security and control. Private and hybrid clouds should be considered by organizations running SAP to add agility and flexibility to their landscapes. Hybrid configurations allow certain projects to leverage internal architecture while others can scale to public cloud systems after certain decision criteria are evaluated such as cost, capacity, performance, turnaround time, how long the environment is required, and data sensitivity. Organizations can start preparing for a hybrid cloud by performing an extensive review of their current infrastructure costs and capabilities. Internal costs for deploying and maintaining SAP systems should be determined and compared to public and hybrid cloud options. It is also important to understand how long it takes to deploy a system internally versus public and hybrid clouds and what system performance levels can be achieved in both scenarios. This information, coupled with automation, creates a more effective and efficient provisioning mechanism when compared to traditional methods. SAP customers should evaluate cloud computing technologies in the near term to determine what benefits may be useful to their organization. While there are specific benefits for SAP landscapes, cloud computing IaaS should be evaluated as part of the organization’s overall IT strategy including a full ROI analysis. As the technology evolves, operational experience is gained, and standards emerge, there will be increased opportunities for SAP customers to leverage features of cloud computing.