Cloud Computing: Myths and Realities Viewpoint paper the real value of Cloud to your organization The IT marketplace can t resist giving new names to technology services and the ways in which they re delivered. Familiar examples of this include shared, electronic data interchange (EDI), leveraged, web, utility, grid, and software as a service (SaaS). In some cases, the new names reflect meaningful distinctions and advancements. In others, some argue, they re the same wine in a new bottle.
Table of contents Cloud Computing: Myths and Realities...1 Myth 1:...2 Myth 2:...3 Myth 3:...3 Myth 4:...4 About the authors...5
In the lexicon of IT terms, cloud computing has now taken center stage. Technically, the use of cloud computing dates back at least a decade in the form of application hosting, which many now refer to as SaaS. Cloud Computing: Myths and Realities Much of today s attention centers on public clouds, which are similar to utility-like services that can be tapped into via the Internet on a pay-per-use, or in some cases, free basis (Figure 1). These services are seen as a way to reduce capital costs or overcome a lack of internal IT capabilities, while still having ready access to computing resources whenever needed. But such services are not a panacea. And there is understandable wariness among large businesses that require enterprise-class IT resources and services about what public clouds can and cannot deliver. Because of this, many businesses are also investigating private clouds, or utility services. While not entirely the same, these service solutions are configured similarly. Both are predefined, prebuilt, and energyefficient standard environments provided as a service through Information Technology Infrastructure Library (ITIL)-based processes. Figure 1 Next-generation services LOW Outsourcing Delivery Standardization Private Cloud Utility Services HIGH Public Cloud CORE INFRASTRUCTURE Server/Infrastructure Services Storage WORKPLACE SERVICES Messaging Collaboration Security, WP as-a-service WEEKS Time to Provision HOURS 1
Delivered from professionally managed, state-of-the-art global data centers, private cloud services offer choice and flexibility in how IT services can be deployed and help speed infrastructure and application implementation, while reducing risk and cost. Because all assets are owned and managed by the service provider, these services allow businesses to convert the capital investment associated with an infrastructure build-out into a monthly operating expense. Leading private cloud providers are offering services through converged infrastructure architectures. These environments employ service automation and orchestration to speed changes, reduce errors, and strengthen compliance. They capitalize on the increasing use of virtualization, while supporting nonvirtualized operations. Components of a converged infrastructure include operating software that manages heterogeneous environments; a flexible, wire-once network fabric that allows for multiple protocols and changes without incurring downtime; virtualized server, network, and storage pools that support allocation of resources where they re needed; and a data center smart grid for effectively power and cooling management. HP is helping organizations investigate cloud computing and develop strategies to capitalize on its potential. In this viewpoint, we explore some of the myths and realities of cloud computing. Understanding these issues can help businesses make sounder decisions on provisioning software and services to address business objectives. Myth 1: I will always save money with cloud computing. Reality: I will save money if cloud computing fits my business requirements. Organizations often investigate cloud computing in the hope they will be able to free themselves of the expense of owning and maintaining infrastructure. They can exchange capital costs for operational costs. But it is not as simple as that. There are costs associated with compute resources not meeting expected performance levels, not being available when needed, and not being sufficiently secure to protect enterprise and customer information. Also, while IT resource consumption is flexible in a cloud environment, the contract underlying those services may not be as flexible as expected. Certainly there is value in adopting the technologies that go into creating a cloud environment, such as virtualization and automation. But cloud environments may or may not be a good fit for organizations with business-critical IT requirements. It can be helpful to consider a provider that offers both cloud and more traditional hosting services. Such a provider can create the most appropriate solution for particular needs. A first step in deciding whether and how to use cloud services is to examine your various IT requirements and assess which may or may not be good candidates. 2
Myth 2: Cloud computing helps me get work done much faster and provides quicker response times. Reality: Cloud computing solutions that employ low-performing assets can yield variable performance, with inherent contention for time on those assets. Public cloud providers typically use blade servers built from off-the-shelf commodity components, which the providers often assemble themselves. They deploy these white-box resources using a scale-out model in which units are added to the system, as needed, to support distributed software applications and other requirements. They have written their own software, which tolerates units dying and the resulting need to hot-swap units. For the most part, the software recovers the system without service interruption. However, problems do arise because the software can only handle limited variations of events that can occur. As a result, public clouds at their core are not high-performance computing resources. They are inexpensive and have connectivity, but they are not likely to provide fast transaction times or the ability to perform analytics. They often cannot accommodate storage demands, response times requiring high levels of caching, or extensive short-term memory capacity needs. In contrast, enterprise-class private cloud services provide performance-enhancing capabilities including monitoring, redundancy, and disaster recovery that are typically not found in public clouds. Myth 3: Cloud computing will provide access anytime, for any application. Reality: Cloud computing can provide access within guaranteed limits. The ability to access a system when needed and have it deliver the results required is a function of availability and reliability. Availability means a system is always there when accessed. Reliability means it provides the results expected from it. In deciding the importance of availability and reliability, it is useful to keep in mind the adage that time is money. Inexpensively built systems such as those deployed in public clouds are susceptible to downtime, which could result in users not being able to access your website, or leave you unable to query your enterprise resource planning system. For companies whose IT infrastructure supports business-critical functions, seconds of downtime can result in losses of hundreds of thousands of dollars. In retailing, for example, hours of downtime can cost millions of dollars, especially during certain times of the year, such as Black Friday or the Friday after the U.S. holiday of Thanksgiving. Across industries, a loss of system availability on December 31 when a company is trying to close out its books could result in losses affecting the entire fiscal year. Scale-out systems, such as those employed by cloud providers, are built to fail and tolerate failure. And for the most part, they can meet the needs of small businesses and nonbusiness-critical uses, such as test and development. 3
In contrast, private cloud service providers deploy systems built with higher-quality designs and components that not only provide software failover capacity, but build it into the hardware itself. They are built to keep running, and if something stops running, extensive backup capabilities ensure there is minimal downtime, if any at all. The difference between the two environments is important to keep in mind when considering whether to extend an application into a cloud environment. Further, many enterprise applications are simply not designed to be cloud enabled or even for use in highavailability clusters. For businesses considering cloud deployment of critical applications, it is essential to determine if the service provider will offer and abide by a service level agreement, and at what cost. With enough time and effort, virtually anything can be done with technology. The question is who is paying for it? Myth 4: Cloud computing provides ample security and is low risk. Reality: Cloud computing can comprise an unknown set of assets in an unknown location with unknown operators. International business creates particular challenges, as different jurisdictions have differing, and in some cases, highly restrictive regulations governing what information on their citizens or financial data can cross borders. When assessing potential cloud services, it may be impossible to determine where systems and the people operating them are located, or how secure those systems are. Because of this, cloud operations may not meet regulatory requirements or the demands of cross-border commerce. Businesses will continue to explore the potential of cloud computing in their quest for the perfect elasticity of indefinite delivery, indefinite quantity the Holy Grail of IT services. As these myths and realities suggest, there are situations in which the technology is not ready for the heavy lifting of enterprise-class IT. In such cases, turning to a service provider that offers private cloud services delivered on an efficient, economical, converged infrastructure may be a promising alternative approach. By carefully evaluating the potential benefits and pitfalls of cloud computing, businesses can deploy it for uses that make sense, while avoiding the problems of overestimating its capabilities. Risk takes many forms from financial exposure in transactions to defense and national security issues to public safety. Downtime can be a million-dollara-minute problem for some businesses. Government mandates for record retention require businesses to have disaster recovery capabilities. Development of products such as pharmaceuticals requires thorough tracking of product pedigrees. 4
About the authors Bennett Bauer Bennett Bauer is a Product Marketing manager for Application Hosting, Web, and Utility Services offerings at Hewlett-Packard. Bauer was previously engaged in marketing high-end HP servers and software, as well as innovations in IT as a service model. Before HP, Bauer held operations and marketing positions with emerging to mid-sized technology companies in the areas of supercomputing class servers, storage I/O, and IT security products. Bauer holds an MBA from Santa Clara University and an industrial science degree from San Jose State University. A.J. Murray A.J. Murray is responsible for Marketing and Sales Readiness for Application Hosting, Web, and Utility Services offerings at Hewlett-Packard. Murray s previous roles at HP include that of a Web Hosting Consultant and storage delivery support. Before HP, Murray held various sales and support positions in the healthcare software industry and in telecommunications companies. Murray holds a masters degree in international relations from the University of Denver. 5
Share with colleagues Copyright 2010 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein. 4AA0-1810ENW, Created May 2010