Research Publication Date: 27 September 2007 ID Number: G00148919 Alternative Delivery Models: The First Taxonomy James Holincheck, Claudio Da Rold, Mark A. Margevicius, Bruce Robertson Today's environment allows for a decoupling and modularization of IT, thereby changing the definition of IT as well as the way IT is delivered to the business. There is a spectrum of choices that will flourish between these extremes: Building and managing all IT in-house Acquiring all IT capabilities as a service As more options appear, the lines between them will blur, offering organizations a continuum of choices. In this research, we present a first taxonomy of the new and alternative IT delivery models. Key Findings Alternative delivery models are emerging to enable the delivery of different levels of IT service: user devices (PC, mobile), infrastructure (network, data center, storage, computing, grid), application (streaming, software as a service [SaaS]) and business processes/services. Alternative delivery models are, therefore, emerging at and affecting different layers of IT architectures, ranging from the technical component layers (computing utility, CPU on demand) and technical domains (network, security) into technical pattern (grid) and technical services that can be further composed. Recommendations Clients should begin the process of evaluating alternative delivery models by: Understanding the differences between services provided through a new delivery model and a traditional delivery model Evaluating whether new solutions and services emerging in the market can deliver business value or solve IT issues better, faster and for a lower cost than traditional approaches Understanding how these new models will affect the enterprise architecture, delivery management processes, the sourcing strategy, and the IT organization duties and responsibilities Reproduction and distribution of this publication in any form without prior written permission is forbidden. The information contained herein has been obtained from sources believed to be reliable. Gartner disclaims all warranties as to the accuracy, completeness or adequacy of such information. Although Gartner's research may discuss legal issues related to the information technology business, Gartner does not provide legal advice or services and its research should not be construed or used as such. Gartner shall have no liability for errors, omissions or inadequacies in the information contained herein or for interpretations thereof. The opinions expressed herein are subject to change without notice.
ANALYSIS In "Alternative Delivery Models: 14 New Ways to Deliver IT," we define a first list of the major new and innovative delivery models that are starting to populate the IT market. Also, a list of models that are considered "traditional" for the scope of this research is provided. This research provides definitions for those delivery models for which a Gartner or market definition is available. In this research, we present a first taxonomy of models, based on the "Relationship/Implementation Matrix" for the delivery of innovative services to business organizations. These models have significant degrees of flexibility/variability, and some of the new models are still in their early days of development. This represents a challenge for a complete taxonomy, because each model may have different flavors. When comparing different models (or a single offering belonging to one of the models) as part of their sourcing strategy or solution selection, client organizations will be able to better specify the type of offering or model they are evaluating. In the future, we will continue to refine this taxonomy as models and offerings evolve. Structure of the Taxonomy There are many models discussed in "Alternative Delivery Models." The first step in creating the taxonomy was to recognize that alternative delivery models are just that, alternatives. They provide another way to go than established or traditional delivery models of application or infrastructure delivery to a certain type of buyer. Thus, the taxonomy focuses on comparing alternative and traditional delivery models against one another. This taxonomy does not include all alternative or traditional delivery models. We used the concept of viewpoints to group logically related delivery models together. A viewpoint distinguishes a particular buying audience or common sets of applications/technologies. Finally, we identified for each viewpoint the relationship and implementation criteria that define each offering so that a buyer with that viewpoint will be able to see the differences between alternative and traditional delivery models. We have highlighted the alternative delivery models in green in Figure 1. Publication Date: 27 September 2007/ID Number: G00148919 Page 2 of 11
Figure 1. Business Process/Applications Viewpoint of Alternative Delivery Models On-Premise Application Maintenance Outsourcing Application Hosting SaaS BPO BPU Relationship Responsibility for Process Responsibility for Technology/Support - Application Layer / - Infrastructure Platform Implementation Location of Application Processing * * * * Single Line of Code and Data Definitions Support Multiple s Financial Model Purchase + Maintenance or Lease Fixed Fee or Time and Materials Purchase + Maintenance + Hosting or Pay per Use Many or Pay per Use *Client-side processing would be allowed. BPO = business process outsourcing BPU = business process utility Source: Gartner (September 2007) Publication Date: 27 September 2007/ID Number: G00148919 Page 3 of 11
On-Premise From a business applications/process perspective, this is the most familiar model. The customer is responsible for the business process, the technology and technology support (for the applications and the infrastructure platform). The application is installed at the customer site (each customer implements its own instance of the code and data definitions). The common financial model is a perpetual license with annual maintenance fees. A subscription model typically is not used, but some vendors offer leasing options and financing (so it looks like a subscription). Implementation of ERP solutions from vendors such as Oracle and SAP are good examples of on-premise delivery. Application Management Outsourcing Application outsourcing refers to a multilayer contract/relationship involving the purchase of ongoing application services from an external service provider (ESP) that supplies the people, processes, tools and methodologies for managing, enhancing, maintaining and supporting custom and packaged software applications, including hosted applications. Application outsourcing contracts may comprise a broad portfolio of services, such as application development, integration, deployment and management services, as well as consulting/advisory services. Discrete development projects or staff augmentation services are not considered application outsourcing. Application outsourcing services may be delivered on-site, onshore, nearshore or offshore. They may also be delivered in traditional ways (labor-based on-site or off-site) or via a hosted or usage-based delivery model (see "How to Make Better Application Outsourcing Decisions"). Large system integrators such as IBM and Accenture, as well as offshore firms such as Tata, Wipro and Hexaware provide these services (see "Hype Cycle for IT Outsourcing, 2007"). Web/Application Hosting Web/application hosting services can be thought of as a "stack" of services. These hosting vendors provide management services at some or all of these layers (they may partner with other providers for some of the layers). The layers encompass data center facilities, Internet and WAN connectivity, including routing, local-area networking (including switches, firewalls, load balancers, Web caches and other specialized equipment), server hardware, storage (including network-attached storage and storage area networks, operating system, Web server, application server, other middleware and packaged enterprise applications, and database. Many different providers offer application hosting services, including Data Return and NaviSite (see "Hype Cycle for IT Outsourcing, 2007"). There are four main differences between this model and application management outsourcing. The first difference is in the service. In Web/application hosting, the service provider is running the application, while in application management outsourcing, it is managing the application code. Second, the location of the application is at the provider's site, not the customer's. Third, in the hosting model, the customer or the provider may be responsible for the application layer (it is typically the customer). Fourth, the financial model is different. Application maintenance outsourcing typically is priced based on a fixed fee or time-and-materials basis, while application hosting is usually based on the number of users or servers managed (see "Market Overview: Web Hosting, Application Hosting and Utility Computing"). Software as a Service (SaaS) The SaaS model is similar to the application hosting model in that the provider owns, manages and operates the application remotely. However, the SaaS model differs from application hosting in two important ways. First, SaaS requires that a single set of code and data definitions is used across multiple customers. Second, the licensing model for SaaS is either subscription or pay-per-use (see "Findings From the Gartner 2006 Global Research Meeting: Enterprise Application on Demand" and "How to Evaluate SaaS Architecture Model Choices"). Many vendors offer SaaS business applications, including salesforce.com, NetSuite and Taleo (see "Hype Cycle for Software as a Service, 2007"). Publication Date: 27 September 2007/ID Number: G00148919 Page 4 of 11
Business Process Outsourcing We define BPO as the delegation of an IT-enabled business process to a third party that owns, administers and manages the process according to a defined set of metrics. Specific BPO offerings include: Horizontal offerings (those that can be leveraged across specific industries), such as benefits administration, accounts payable, indirect procurement, contact center operations and more-holistic horizontal offerings, including HR-related functions. Example providers of horizontal BPO services include Hewitt (HR BPO) and Genpact (F&A BPO). Vertical-specific offerings, such as revenue cycle management for the healthcare industry, mortgage origination for the banking industry and policy administration for insurance companies. Billing Services Group (billing for telecom providers) would be an example of a vertical-specific BPO provider (see "Hype Cycle for Business Process Outsourcing, 2007"). In the previous delivery models, the customer was responsible for the business process (managing the inputs, the processing and the outputs). With BPO, the provider becomes responsible for the process. In addition, BPO is similar to application hosting in that the service provider typically owns, manages and administers the technology at its site and implements a separate instance of code and data definitions for each customer. Finally, it is similar to SaaS in terms of the financial model, although there are many types of financial relationships possible in a BPO deal. Business Process Utility Much like in BPO, a BPU would be responsible for the business process and the technology/support. However, like a SaaS provider, the BPU uses a common set of code and data definitions consumed by customers in a one-to-many model. In addition, it charges on a subscription/pay-per-use basis. Vendors such as PayPal and Exigen are examples of emerging BPUs (see "Hype Cycle for Business Process Outsourcing, 2007"). These are not "one size fits all" models. s should expect to use all these different models to enable their business processes (see Figure 2). Publication Date: 27 September 2007/ID Number: G00148919 Page 5 of 11
Figure 2. IT Infrastructure Viewpoint of Alternative Delivery Models On-Premise Infrastructure Remote Management Services Hosting/ Colocation Services Grid Computing Infrastructure Outsourcing Infrastructure Utility Relationship Responsibility for Technology Architecture Design / Responsibility for Technology/Support Implementation Location of Technology / or Single Infrastructure Platform Supports Multiple s Financial Model Purchase + Maintenance FTE-based Fees or Volume-Based (number of servers) Purchase and Fixed Price or Baseline/ Volume-Based Baseline/ Volume and Usage-Based Source: Gartner (September 2007) On-Premise Infrastructure This is the traditional model where the customer is responsible for the design of its IT architecture (hardware, system software, storage and network everything but the applications). Remote Management Services This is similar to on-premise hardware, except that the service provider is responsible for supporting the hardware/devices, and the financial model is recurring revenue-based (see "Forecast: Remote IT Monitoring and Management Services, Worldwide, 2005-2010" for the definition and more details about the market). Avaya, Getronics and HCL are examples of remote management services providers specifically for network management (see "Hype Cycle for IT Outsourcing, 2007"). Hosting/Colocation Services Colocation is the provisioning of data center facilities, together with Internet and private WAN connectivity. Additional basic services, such as monitoring, remote hands and backup tape rotation, may be offered. This segment excludes more-complex server management; colocation customers will do this themselves or outsource to a third party. In this model, the customer is responsible for the hardware architecture and design, but the service provider manages the facility where the hardware will be located and supported. There is usually a monthly usage fee associated with the facility (see "MarketScope: Internet Data Center Colocation, United States, 2005"). AT&T and Savvis are good examples of providers of Internet data center colocation services (see "Hype Cycle for IT Outsourcing, 2007"). Grid Computing Grids are collections of computer resources, owned by multiple organizations, that are coordinated to solve a common problem. These resources can be computers, collectively run in extremely large parallel-processing programs typically used to solve the large-scale problems found in scientific and engineering computing. The multiple organizations may be product or geographic divisions of one firm or may be multiple companies Publication Date: 27 September 2007/ID Number: G00148919 Page 6 of 11
(see "What Grid Computing Is Really About" and "Answers to Questions on Grid and Clusters From Gartner's Symposium/ITxpo 2006"). Grids can be hosted by a third party or they can be run on-premise. Grids can also be provided as a service (where the provider defines the architecture). Vendors such as Appistry, DataSynapse and Platform Computing offer grid computing-enabling technologies (see "Hype Cycle for Server Technologies, 2007"). Infrastructure Outsourcing In IT infrastructure outsourcing, the provider takes responsibility for the ongoing management and operations of data center/host, desktop/client and network environments; help desk; and other infrastructure support. Many of the large, global system integrators provide infrastructure outsourcing services, including CSC, EDS and IBM (see "Key Issues for Outsourcing, Infrastructure and Applications, 2007" and "Hype Cycle for IT Outsourcing, 2007"). Infrastructure Utility (IU) IU is a shared IT infrastructure architecture provided through ondemand services, with pricing based on service use, and proven ongoing reductions in fixed baseline (or subscription fees) and unit costs. The IU is open, flexible, pre-designed and standardized, virtualized and highly automated, highly secure and reliable. IU services are accessible through pre-defined methods, standards, and terms and conditions (see "Infrastructure Utility in Practice: Offerings Description"). Many of the large system integrators have also started to offer IU services, including Atos Origin, BT Global Services, Capgemini and HP (see "Hype Cycle for IT Outsourcing, 2007"). It is important to note that these models not only can be used at an overall infrastructure level, but also for different components of infrastructure, such as hardware, storage and networking. Publication Date: 27 September 2007/ID Number: G00148919 Page 7 of 11
Figure 3. End-User Computing Viewpoint of Alternative Service Delivery Models Desktop Applications Hosted Server- Based Computing Services Server- Based Computing Software Streaming SaaS Web Platforms User- Owned Devices Relationship Responsibility for Technology/ Support - Application Layer s/ (s) End User/ - Infrastructure Platform (s) End User Implementation Location of Application Processing * * (s)* N/A Single Line of Code and Data Definitions Support Multiple s N/A Financial Model Purchase Purchase Purchase on a per- User Basis Ad- Supported/ Purchase * Client-side processing would be allowed. Source: Gartner (September 2007) Publication Date: 27 September 2007/ID Number: G00148919 Page 8 of 11
Desktop Applications These are the traditional desktop applications, such as Microsoft Office, that run on the client. The customer is responsible for the PC technology and supporting the applications. They run on the local PC, not at a third-party site. Desktop applications are typically purchased, not subscribed to or leased. Server-Based Computing Server-based computing is a thinner-client computing style that puts thick-client applications (for example, desktop applications) onto the server where they can be managed centrally. Terminal servers are installed at the customer site, which provides thinner client access to those applications. These servers are priced on a named- or concurrent-user basis. Vendors such as Citrix and Microsoft offer terminal servers (see"client Platforms Spotlight: Evaluating if Server-Based Computing Is Right for You," "TCO Comparison of PCs With Server- Based Computing" and"application Streaming Will t Make Server-Based Computing Obsolete"). Hosted Server-Based Computing Services Some providers offer hosted terminal servers that a customer can use to provide thinner-client computing to end users. So, instead of the customer buying and implementing its own terminal servers, the service provider hosts it and charges a per-user, per-month fee (plus setup fee). Vendors such as EPI and RedPlaid offer hosted terminal servers. Software Streaming Software streaming technology is typically implemented at each customer site. A server streams code to client machines (which can also be cached locally at the client). In this way, customers are able to have a single line-of code and data definitions for all of its users (although it is not a single line of code and data definitions across multiple customers). It is priced on a per-user basis. Vendors offering enabling technology for application streaming include AppStream, Citrix and Microsoft (see "Hype Cycle for PC Technologies, 2007" and "Client Computing Architectures: How to Choose From the Many Options"). SaaS Much like in the business process viewpoint, personal productivity applications can be supplied via the SaaS model. It is more common in personal productivity applications delivered via SaaS to see client-side processing (caching), although most of the processing is typically done on the server hosted by the provider. In addition, it conforms to the other principles of SaaS, such as a single line of code supporting customers in a one-to-many model and delivered via subscription or pay-per-use model. Examples of SaaS solutions in this context would include Google GMail (e-mail) and Zoho Office (personal productivity applications). Web Platforms/Mashups There are multiple types of Web platforms that providers and customers can use. From an end-user computing perspective, mashups are an emerging way to take advantage of these platforms. Mashups mix content from multiple sources by using feeds from public application programming interfaces. An example of this might be a Web application that mashes real estate information from a real estate company's Web site with an HTML mapping interface from another site to show the location of area home listings. Mashups may be created by a customer or by a provider. These solutions may be offered through a subscription license, may be ad-supported or free (through an extended beta testing period) (see "Decision Framework: Gartner's Mashup Matrix"). Many vendors provide the ability for end users to create their own mashups, such as Google (Maps) and craigslist (classified ads), as well as vendors that are providing tools that enable mashup creation, such as Applibase and JackBe (see "Hype Cycle for High-Performance Workplace, 2006"). User-Owned Devices All IT is not provisioned by an employer. Workers have more and more of their own personal technologies that they want to use for work and nonwork activities. Personal laptops, cell phones and so on represent an opportunity (lower acquisition costs) and a challenge to IT organizations (how much end users can support themselves vs. requiring corporate IT support). These devices may come loaded with applications which the end user is responsible for Publication Date: 27 September 2007/ID Number: G00148919 Page 9 of 11
supporting. Although when connected to the network, it may also access applications provided by the business (via delivery models such as software streaming) (see "Portable Personality Solutions Enable User Device Independence," "Steps Toward Managed Withdrawal From User Devices at Work" and"global Company Turns to User Responsibility and User-Provided PCs to Cut Costs"). Remember that although these delivery models, especially the alternative delivery models, focus on delivery via ESPs, internal IT organizations should think about how they apply the concepts and best practices of ESPs to their internal delivery IT services and applications. IT organizations need to think in terms of managing a portfolio of IT products and services that are provisioned in a variety of ways. How those products and services are provisioned will change over time as new products, services and delivery models emerge. te 1 Delivery Models, Definition and Sample List IT Delivery Model: A delivery model is the use of business processes, people, information and technology that delivers IT-intensive capabilities to organizations to support specific business outcomes. Traditional Delivery Model: The use of the proven, accepted and established processes, technology, best practices and channels required to acquire and deliver IT. Alternative Delivery Model: The use of nontraditional methods, processes and technologies to acquire and deliver IT functionality (elements of the delivery chain) and business-it services (results and business outcomes) in a flexible and use-based fashion. Examples of Alternative Delivery Models: User-owned devices, software streaming and software-based appliances, Web platforms, community source, remote management services, communication as a service, capacity and storage on demand, grid and utility computing, infrastructure utility, SaaS and business process utility. This research is part of a set of related research pieces. See "Alternative Delivery Models: A Sea of New Opportunities and Threats" for an overview. Publication Date: 27 September 2007/ID Number: G00148919 Page 10 of 11
REGIONAL HEADQUARTERS Corporate Headquarters 56 Top Gallant Road Stamford, CT 06902-7700 U.S.A. +1 203 964 0096 European Headquarters Tamesis The Glanty Egham Surrey, TW20 9AW UNITED KINGDOM +44 1784 431611 Asia/Pacific Headquarters Gartner Australasia Pty. Ltd. Level 9, 141 Walker Street rth Sydney New South Wales 2060 AUSTRALIA +61 2 9459 4600 Japan Headquarters Gartner Japan Ltd. Aobadai Hills, 6F 7-7, Aobadai, 4-chome Meguro-ku, Tokyo 153-0042 JAPAN +81 3 3481 3670 Latin America Headquarters Gartner do Brazil Av. das Nações Unidas, 12551 9 andar World Trade Center 04578-903 São Paulo SP BRAZIL +55 11 3443 1509 Publication Date: 27 September 2007/ID Number: G00148919 Page 11 of 11