Web services have one huge advantage over other IT step-changes, says Neil McEvoy; you don't have to wait forever for the promise to be realized. A smart business can make Web services technology pay for itself in the short term while at the same time securing the long-term strategic future of the enterprise.
There is one vital difference between Web services and other 'mould-breaking' information technologies of recent years; Web services can actually pay their own way.
The adoption of Web services technology offers businesses a massive long-term strategic advantage. But the same technology can also be used to save money and make money in the short term -- in some cases so much so that the technology will have paid for itself long before its strategic long-term benefits have been felt.
As always with a new information technology that has the potential to change the business world, industry commentators are split neatly between those who are hyping it to the heavens and those who advise waiting to see if it's just another passing fad.
The skeptics point out that the full promise of Web services cannot be realized until open standards are fully developed, and the Web becomes secure, fast and reliable enough to support exchange of mission-critical data and applications in an 'on-demand' manner. Until then, they believe, buying into the technology would be little more than an expensive knee-jerk reaction to vendor hype.
Let us be clear; the vision of a futuristic 'service grid' where dynamic machine to machine communication is the backbone of global trade is just that -- futuristic, perhaps years away. But with the right approach, adopting Web services technology right now could yield such a level of savings that a return on investment could be achieved in a matter of months, particularly for large enterprises.
The key to doing this is to broaden the scope of thinking with regards to the underlying technology and where and how it can be applied. Most discussions concentrate on the XML standards, the idea of small applets such as a credit checking service, plugged easily into new applications, and as a slightly improved solution to the labour and cost intensive EAI projects that corporates wrestle with. All well and good, but not likely to get the CFO reaching for his checkbook in a hurry.
However, if we look under the bonnet at the underlying technology engine, we see a platform capable of enabling the 'service-oriented architecture', the term used to describe the on-demand framework for Web services. As its name implies, the service-oriented architecture (SOA) focuses on making resources available as a service, in that they can be consumed immediately but only as and when needed, and without the capital cost and operational hassle of the acquisition, build and set-up phases. The architecture creates an environment in which consistent, open standards are used to define how each resource can be 'plugged in' and service-enabled.
This model and technology, applied internally within the organization, can improve the overall commercial effectiveness of a company's operations at the 'building block' level -- how it defines and executes business processes, and how these processes are able to achieve their objective through manipulating resources without human intervention.
It is this exchange of human effort and capital acquisition for automatic transaction and 'on demand' usage that fuels the vision of Web services extending across the global trading environment -- but they can also be used for very down-to-earth and practical purposes right here and now.
A fairly logical place for the IT team to implement this approach first is the IT environment itself, so that the supply of applications, infrastructure and maintenance is delivered via the SOA, improving service delivery to internal users, increasing performance and resilience, and decreasing costs.
Some simple examples, restricted just to the IT area, reveal areas where savings are available:
- IT administration: Consider how the minutes and hours add up for systems administrators engaged in the routine tasks of setting up and maintaining simple IT resources for company users. It may only take a few minutes to set up an email account, but what about network access, accounts for multiple business systems, numerous password support issues, installing new software, ordering and setting up laptops and so on. Even a minimal estimate of two hours sysadmin work per staff member per year across all these functions would cost a global enterprise with 50,000 employees more than $5m a year in labour costs alone -- and it is my guess that for the typical enterprise the routine IT load per employee is far in excess of the two hours cited above. Yet Web services technology, applied to the sysadmin area, can effectively automate all these repetitive tasks.
- Downtime costs: It is only natural with this level of administrative work being performed by humans that errors occur. Even a simple mistake in setting up an email account or laptop keeps a worker offline for hours, losing productivity. This is difficult to quantify financially, so perhaps instead consider misconfiguration in an area such as DNS. Something as tiny as a comma instead of a full stop when keying in this data has a total and devastating result: complete unavailability of a Web page or even the whole site, during which downtime direct sales revenues are being lost and unknown damage is being done to the brand and reputation of the company.
Automation of operations in these areas removes the need for human involvement, eliminating the original costs, the errors caused, and the resulting knock-on secondary costs. (Consider the fact that DNS lookup failures have been analyzed by Keynote as the second most common reason for Web site failures, accounting for 29 percent of failed Web page downloads. A Gartner analysis estimated that as much as 40 percent of infrastructure downtime is caused by operator error.)
The automation of IT infrastructure maintenance is quickly becoming a hot subset of the service architecture technology market, with start-ups such as Terraspring (terraspring.com), Jareva (jareva.com), Ejasent (ejasent.com) and ThinkDynamics (thinkdynamics.com) specializing in this area, as well as vendors pitching in with major plays, such as Sun's N1 platform and HP's Utility Data-Centre vision.
None of these companies are offering Web services solutions per se, but if we boil them down to their bare bones we can see the same technology model as cited by those talking more about XML and EAI, such as Cape Clear and Iona -- i.e. technology that allows resources to be defined as services available on the network via common interfaces, and a process engine to sweep along and 'orchestrate' them together to fulfil user needs. Instead of business processes exposed from mainframes and other legacy systems, we have data-centre assets, such as storage and application infrastructure, but the model is still the same. This is the Service-Oriented Architecture in effect, just starting with different business objectives in mind.
And remember, this is just thinking about Web services technology as applied to the very narrow field of internal IT administration; a moment's thought will throw up scores of other internal areas (invoice and credit control processing, payroll, training, maintenance schedules, personnel administration, health and safety routines etc) where the technology can be used to achieve concrete savings. Yet this investment -- of such great short-term benefit -- is also creating the same platform that will be able to take advantage of the strategic business opportunities offered by XML Web services in the medium to long term.
Furthermore, the fun doesn't stop here. This 'IT process engine' enables the organization to be more effective at acquiring resource from suppliers, creating the potential to find further savings, particularly with next-generation vendors, such as xSPs who already offer their service in an on-demand manner. Instead of buying in expensive top-down enterprise applications with long install times and white-knuckle learning curves, the IT department can begin to deliver user solutions by working on a bottom-up basis, within a modular framework for finding and consuming functional resource automatically as and when it is needed.
To illustrate this principle, a simple example can be made from looking at basic components, such as servers and Internet bandwidth. Corporate IT departments commonly procure excess capacity in preparation for growth in demand, indeed many operate a policy of ensuring utilisation is kept low, with some reports suggesting that the average server utilisation is a surprising eight per cent, meaning that considerable capital has been deployed to create idle resources. The reason for this is that IT knows that should a sudden peak in demand fail to be satisfied because of a lack of available resource, accusing fingers will immediately be pointed in their direction, so naturally they pre-empt this by acquiring excess capacity -- at excessive cost.
The simple and effective solution to this problem is known as 'utility computing', acquiring the resource only when you need it and paying just for what you use. This is achieved through process automation: systems detect when demand is incoming, and processes are initiated to automatically make the necessary resource available. For example, for its revenue-critical e-business site, a company could agree internet bandwidth deals with three different providers, and connect the physical pipes to their servers with extra capacity available but not consumed as standard. By programming processes that can systematically increase and decrease the available IP bandwidth, then this could be driven dynamically by demand: a marketing e-mail shot hits the spot and customers swarm onto the site within a matter of moments, but as incoming traffic suddenly escalates, monitoring systems detect this and these processes then fire into action, invoking more capacity to meet it. Once the customers (bloated with purchases, we trust) retreat from the site, the services are then automatically wound down again.
This approach also brings inherent resilience. Should one of the providers suffer downtime, then the system would detect the decrease in available capacity, and automatically ratchet up alternative sources compensate, thus ensuring that no disruption to the site's availability is caused.
Applying this model across the spectrum of infrastructure casts a new light on planned expenditure within the IT budget, allowing future purchases to be 'cancelled out'. Eliminating planned expenditure from the budget equals more savings to be factored into the business case. And the long-term business benefits of adopting the service-oriented architecture don't stop here, not by a long shot. As well as allowing the company to be plugged into the 'business grid' of the future and making its products more saleable, the architecture provides the basis for bottom-up self-organization, building a corporation that can adapt dynamically to changing market conditions, while at the same time achieving finely granular levels of real-time management reporting.
A host of benefits… but in the here and now, it is the nitty-gritty of cash saved that makes the most compelling case for the adoption of Web services technology. You know when you persuade the Board to buy into the technology that what you're doing is securing the long-term future of the enterprise but you can sell it to them by pointing to this year's bottom line. The bean counters will love you -- and later on, you can reveal your true identity as the visionary who saved the CEO's bacon by building him the technology of the future when he wasn't looking...
Neil McEvoy, CEO, www.webservices-strategy.com, is a specialist in business and ROI strategy for service oriented architecture technologies.