Get on the Grid: Utility Computing

Utility Computing

Traditional enterprise software product licensing is being replaced with more creative hosted or managed "software as service" models. The latest business model is utility computing and associated pricing. Sometimes referred to as "grid" computing, this on-demand model is a trend that has captured the attention of technology heavyweights including IBM, Hewlett Packard (HP), Oracle, and Sun Microsystems. Functioning on concepts similar to common electric and water bills, utility computing allows customers to purchase processing power and software access as needed and to pay based on how much and how often the software is used. This emerging model may even be attractive to large organizations, given that almost all of the initial hosted applications have targeted small and medium businesses (SMB).

In the past, most first-generation hosted applications were not suitable for delivery over the Internet and certainly not appropriate for SMBs, its target market. Application service providers (ASP) were all but pronounced as a failed business model early in 2000 after a raft of availability, privacy, security, compliance, control, customization, and cost concerns by customers and market observers emerged, resulting in many real-life disasters.

Additionally, ASPs faced challenges, the main one being how to drive long-term costs down while accumulating a solid revenue stream. Earlier ASP offerings were based on using a third-party client/server system, but lacked the economies of scale. One cost inhibitor was the amount of dedicated bandwidth that had to be maintained to support thousands of users. Another challenge ASPs faced was service level agreements (SLAs)—if, for some reason, the ASP lost Internet connectivity, customers would lose connectivity to their outsourced production systems and negative impacts on their internal SLAs would result. To make things worse, many of these "flash in the pan" companies went public before making any profits, when the market and investors were disillusioned with the dot-com bubble.

ASPs also over-promised its ability to host heavily customized and concocted enterprise applications environments, which was not technologically feasible at the time. They often tried to fit a "square peg into a round hole" by hosting applications that were not amenable to hosting at that stage (or not at all). Such failures, coupled with the mass demise of early ASPs, and subsequent disgrace of an otherwise great idea from a positively dull, reliable, service-oriented business model ensued. For more information, see Hosting Horrors!.

Revising ASP Value Propositions

However, the current, difficult economy is making customers—particularly those with multiple, dispersed worldwide locations and with many outsourcing partners—prefer this type of subscription model. Because capital budgets have been slashed, everyone wants to avoid one, up-front lump payment, preferring to stretch financing over a prolonged period. Also, users' mindsets have changed significantly over the years. Now multinationals are increasingly sending some operations, like taxation, to third-party companies for processing. Activities that are considered commodities rather than core competencies will likely be the first to go the hosting or subscription route. Yet, given the unsustainable financial model of many early ASPs and due to the many hosting snafus of the past, the move from software-as-a-standalone or pre-packaged, on-premise solution to software-as-a-service will not happen overnight. Nonetheless, because of this growing demand, hosting will get another breath of fresh air in a few years' time, as gigantic IT providers such as IBM, EDS, Microsoft, Oracle, and Accenture will likely compete in the hosting/ASPs market.

Also carrying the "on-demand" moniker, and with more standardized service delivery systems, IBM's recent purchase of Corio, is possibly the best indication of a revised ASP value proposition. In January, IBM announced it was bolstering its presence in the SMB market by paying $182 million (USD) in cash—about twice of Corio's annual revenues—for the APS pioneer. Founded in 1998, and backed by venture capital heavyweight Kleiner Perkins Caufield & Byers, Corio has delivered enterprise applications over the Internet since its inception. Today, the San Carlos, California (US) hosting provider delivers applications from vendors such as Ariba, Oracle (including former PeopleSoft), SAP, and Siebel Systems, many of which have been IBM enterprise applications partners. The purchase also gives IBM five new data center facilities, including one in the heart of Silicon Valley.

Yet, despite its longtime role in the ASP market, the 300-person Corio has shown only modest success, with five consecutive years of losses. Undoubtedly, IBM is not spending a hefty amount of money for Corio's existing revenue stream, but rather for its know-how of scalable hosted applications. The ASP provider has espoused rapid deployment capabilities, an application deployment and management platform, and configuration tools that other ASPs have never invented. Essentially, IBM is buying Corio's tools and people, to complement them with the scale of IBM Global Services (IGS).

This is Part Two in the Trends in Delivery and Pricing Models for Enterprise Applications series.

Key to ASP Success

Meanwhile, ASPs that have thrived have done so by first focusing on one or two core areas, such as sales force automation (SFA), e-mail management, e-commerce storefronts and product catalogs, HR/payroll or financial software, and by offering only their own software. Software is written from ground up entirely for an Internet-based architecture and runs on shared servers to minimize costs. Notably, this type of multi-tenant installation which combines the data from several companies into one application makes upgrading faster and easier, but also opens the door to cross-company errors in data security, corruption, or compromise.

Still, the key to an ASP's success will continue to lie in the targeted marketplace. ASPs that can successfully market to SMEs while providing industry-specific focus and good technical support coupled with frequent software and hardware upgrades will be successful. These upgrades will also need to be transparent and non-disruptive to the user. Customers must be confident that the provider is doing everything right with its product architecture, security, and choice of partners. In addition to these concerns, prospective user companies looking at software as a service should also consider the provider's long-term viability. Additionally, some hosting providers may also provide both software and hosting services. However, the loss of such a provider will mean the loss of both services and this single point of failure will be a huge risk for the user.

Also, as customers typically fancy the liberty to switch from hosted to internal deployment, many large vendors, including SAP, Oracle/PeopleSoft, Microsoft, Best Software, and Siebel Systems, are pursuing a dual strategy that allows both on-premise and hosted systems. For example, Oracle has been particularly aggressive in promoting its ASP/hosting service, called Oracle OnDemand. For a maximum price of $150 (USD) per user per month, it allows smaller companies to deploy either parts or the entire Oracle E-Business Suite, which is hosted in Oracle's own data centers. Customers still have to pay Oracle's software license fee, but the hosting service relieves the customer of providing the infrastructure and IT staff to support the system.

In addition to deployment flexibility, which allows customers to select what applications will be run on premises or be hosted, it is also advantageous when the vendor is also the ASP. While many tier two vendors provide a hosting alternative, they typically offer the service through third ASP parties. That often means some additional licensing, implementation, or service and support intricacies to the end user. This could make a contact-winning difference to some customers who prefer to deal solely with one vendor.

Further, software as service/hosting models are still evolving and many technological problems that spoiled the first wave of ASPs have been (or are being) solved. For instance, single software instances of the software as a service model typically makes the integration of disparate applications difficult. Conversely, with traditional, on-premise licensing business models, customers can feasibly modify the local instance of the software application to interface with other legacy applications. This has not been quite possible with single-instance hosted services.

Web Services Still in the Future

In order to solve the problem of customizing and integrating to existing legacy applications, vendors and ASPs have also been looking to Web services. However, it will be some time before the technology is mature enough to replace the traditional, pesky message brokers and integration hubs that major companies depend on. For now, these are still the better for high-volume processing than Web services.

The development of technologies like Web services and extensible markup language (XML) transactions standards is making multivendor application interfacing more feasible because hosting companies will presumably be able to develop standard interfaces for application integration. However, at this stage, Web services only have a tiny share of the integration pie. Successful vendors and ASPs must still be able to provide a complete gamut that would support legacy messaging protocols as well as Web services.

Vendors should strive towards providing not only customization, the integration of various hosted services before they touch a customer's enterprise. For more information, see Understanding SOA, Web Services, BPM, BPEL, and More. For the time being, the hosted software service business model still applies best to applications that can be run in isolation or with limited interfaces to third-party applications.

In addition to improvements in security and the reduced prices of bandwidth, hosted software finally looks and feels more like traditional "rich UI" client/server applications. It uses tools like JavaScript to provide error-checking and validation at the client side and perhaps more importantly, the development of dynamic hypertext markup language (DHTML). DHTML is an important feature because it allows Web content to change each time it is viewed. It enables a Web page to react to user input without sending requests to the Web server. This, in particular, has introduced a degree of interactivity to web-based applications that was simply not possible a few years ago. At the time, users were either forced to fill in and submit cumbersome HTML forms, or they had to run additional thin client software, such as Citrix MetaFrame, on their desktops. Neither of these approaches was satisfactory and, while thin client software promised a good solution, it still had a number of shortcomings.

First, thin client software increased costs because it had to be licensed on a per-user basis, (which was meant to be done away with), often causing the cost of the Citrix MetaFrame license to account for more than half of the less-than-enticing monthly fee. Second, because the screen was effectively being "rendered" miles away from user's location, the approach was slow and increased the network load at a time when bandwidth was at a premium price. Finally, users still had to run and maintain the proprietary thin client software on their desktops, defeating the object of running server-centric, web-based software. Ultimately, costs were not being cut by as much as promised. Currently, the need for complex client side interactions has been mitigated with the help of DHTML, which minimizes server round-trips and page refreshes. Yet now the browser's challenge is the lack of support for very advanced graphics. Nonetheless, one should expect more developments in the foreseeable future that will enable more sophisticated interaction scenarios.

New advances in identity management, which provides the unique name of a person, device, or the combination of both that is recognized by a system, are bridging user provisioning between the hosting provider and user enterprise. For example, a federated network identity scheme has multiple identity repositories instead of a central logical identity repository. Identities that are recognized can be enabled throughout the Internet or the local area network (LAN) of an enterprise. This is important to enterprises managing the access individuals to certain resources, including applications, information, and even equipment. Still, properly integrating identity and security infrastructures among customers and hosted services remains a challenge, albeit various technologies (some of which have been developed by a handful of ASP survivors) are making hosted provider platforms more reliable, scalable, and secure. Using software as a service prompts several other customer concerns about integration, whereby the chief concern is data security. Other concerns surrounding software services are system security, back up and recovery, and disaster recovery. Vendors are addressing these concerns in the same way that ASPs have addressed them: with ever more robust equipment, policies, and procedures.

Major Hosting Advantages

To recap, major potential hosting advantages can be

  • More predictable, fixed costs for customers. With the on-demand setup, users will pay only for software actually used, which changes with variable demand.

  • Reduced setup and configuration time, and greater operational simplicity.

  • All upgrades applied to the hosting provider/ASP site-based servers. No need for client or desktop upgrades.

  • Limited funds required for initial startup.

  • Reduced need for internal IT support.

  • Enterprise applications maintenance performed automatically by external experts.

In addition to some aforementioned concerns and issues, topping the list of doubts is whether hosted offerings can be properly integrated with existing on-premise applications. Also there is skepticism about the usefulness of adapting hosted/on-demand solutions to unique business processes and practices. Enterprises want more control of their applications, as they need to be constantly changing configurations, thereby adding new products, developing closer integration between their systems and introducing "best-in-class" business processes. Also, many territorial IT managers will not be pleased to relinquish parts of their IT kingdoms, and relying on an outside host's ability to run their datacenters, even if the host is a viable business. These plus the lack of overall awareness may be why some recent studies have found that over 60 percent of enterprises currently prefer the perpetual licensing model on-premise over subscription-based options. What may be surprising is that the findings show that more companies have outsourced their SFA/CRM, payroll and tax compliance operations than those that have outsourced their accounting departments. This calls into question whether customer information has more corporate importance than information about accounts payable and receivable and if the former should be protected on-premise (which will become a moot point, if one loses its customers).

This concludes Part Two of a four-part note.

Part One defined pricing options.

Part Three will detail the effect of the transition on vendors.

Part Four will cover software as a service business model and make user recommendations.

comments powered by Disqus