![]() |
![]() |
||
|
|
![]() |
CBDI Web Services Roadmap - Guiding the Transition to Web Service and SOA Sponsored by
Sponsored PapersWeb Services Roadmap for the On Demand Business - IBM Vendor Web Services Roadmap Report - IBM. IBM's strategy today is centered around "Business On Deman... more Service Oriented Architecture. An Introduction for Managers Many Organizations are now undertaking development of service oriented architectures, but the probab... more Modernizing Application Integration with SOA Whilst investment in Application Integration initiatives over the last decade has undoubtedly improv... more |
This report has been rated by Forum Members on a scale of 1 to 5 as 4.4 : ![]() ROI - The Costs and Benefits of Web Services and Service Oriented ArchitectureHaving difficulty persuading your colleagues or business sponsors of the merit of adopting Service Oriented Architecture (SOA) or using Web Services instead of some existing mature technology? Need to understand what the costs will be, not just the potential benefits? Then this section should help.
Why Web Services?After a couple of years of hype you would imagine everyone is now familiar with the justification for using Web Services. And yet, on many occasions the industry still struggles to clearly put across the benefits of Web Services and articulate precisely how they differ from existing technology solutions. The issue is that superficially many of the benefits that are attributed to Web Services have also been claimed by pretty much every new technology over the past n years. You might realistically argue that the IT industry is well known for hyperbole and exaggeration, and that reality generally falls short of expectations. Things like "improving business agility", "reducing time to market", etc are still valid - but not entirely new. Why is Web Services going to deliver this time? Consequently, there is a real need to be much more precise about the specific cost savings and benefits for both business and IT that can reasonably be attributed to Web Services. This is not so surprising because Web Services have been a technology led paradigm, and early usage has often been in the area of internal integration where benefits are quite straightforward. Grand visions of everything being connected in dynamic real time scenarios are one thing, but most organizations have more mundane problems to solve. And so we seemingly get stuck in the middle, not always knowing which aspect of Web Services to highlight. Too visionary and the audience gets scared (early adoption = risk), too mundane and they get bored (their existing solutions suffice) Also, the constant evolution of Web Services means that the benefits keep evolving too. What started out as a simple distributed computing solution can now be applied to wider range of connectivity scenarios. So just what are Web Services now? Is it just a better, cheaper, faster (BCF) interface for existing EAI, distributed computing, EDI or other scenarios? Or is it more than that? Of course there should be every reason to use Web Services if they are truly BCF in these scenarios, but this is largely an IT centric message. And a frequent response is that existing solutions in these scenarios are mature and largely get the job done - so why convert to Web Services? Perhaps because the ideas are more abstract, we often see the industry failing to put across what are some of the key differentiators of Web Services. For example the richness of the service specification that can be conveyed in WSDL and emerging business process orchestration standards that enable self describing services and self discovering applications. If you just need your developer to connect A to B, then EAI might suffice. But if you want A to dynamically find B, and later switch to C, without developer intervention, and regardless of the technology that A, B and C use, then Web Services is the only solution. So, what then are the benefits and motivations for using Web Services? And what are the precise features that deliver them? Original definitions of Web Services are now outdated as they were too narrow in their viewpoint, which was typically focused on an RPC centric, object/component oriented view of the world. So we offer up a broader definition, which also now recognises that Web Services are not necessarily anything to do with the web: "Web Services provide a simplified mechanism to connect applications regardless of the technology or devices they use, or their location. They are based on industry standard protocols with universal vendor support that can leverage the internet for low cost communications, as well as other transport mechanisms. The loosely coupled messaging approach supports multiple connectivity and information sharing scenarios via services that are self describing and can be automatically discovered." To understand how both businesses and IT departments benefit from Web Services in detail, let's dissect this statement and consider it line by line. 1. A Simplified Mechanism To Connect Applications Regardless Of The Technology Or Devices They Use, Or Their LocationWeb Services will be ubiquitous. Applications running anywhere, on any technology or device will have a Web Services capability available to them. As such, the applications of customers and business partners will be able to participate in an organizations business process, in real time.
2. Based on Industry Standard Protocols with universal supportPrevious solutions are typically proprietary, and even if so called standards have failed to reach universal adoption.
3. Leverages the Internet for low cost communicationsThe high cost of private networks, coupled with the cost of proprietary EDI/B2B solutions has been a barrier to entry for many organizations. Though big participants can justify this, many industries have thousands of small organizations who cannot. Similarly organizations like retailers with an extensive branch network, can more easily enable them to participate in real time communications.
4. As well as other transport mechanismsWhilst the Internet provides a ubiquitous, low cost transport for Web Services, the "web" in Web Service is now misleading, as bindings to other transport mechanisms are available which may be more appropriate to internal usage or private networks where higher speed and more robust connections are available. Though this might seem to contradict point 3, it nevertheless reflects that for some scenarios the Internet will not be ideal, and that Web Service protocols that will improve the reliability of Internet based communications are not yet mature. The key benefit, primarily to IT, is that it now provides choice of transport.
5. Loosely CoupledPrevious connectivity approaches required the same technology at each end of the wire. For example, even though EAI adaptors enabled different applications to connect to each other, it still required the same proprietary EAI technology as a wrapper around each application. Focusing on XML protocols, Web Services describe the connection, not the technology at either end. Loose coupling is not just a technology issue however, but a key aspect of service design.
6. Supports Multiple Connectivity and Information Sharing ScenariosToday organizations use different technologies for distributed computing, EAI, EDI, B2B, Websites, Portals. This results in n times the products, tools, skills and cost. Web Services provides an opportunity to radically reduce this by supporting these different scenarios with the same basic protocol stack.
7. Self DescribingThe time taken for developers to properly understand how to use an existing interface particularly when it is external to their own projects slows down the time that new connections can be established. Web Services provides a much richer specification of the service compared to previous technologies, which can be accessed programmatically.
8. Automated DiscoveryProvides a mechanism for discovering Service Providers, which can be automated.
Why SOA?Web Services are not a silver bullet. Like most technologies, it is only by ensuring that business requirements are properly understood and their application is carefully designed, that the benefits claimed are truly delivered. It is very easy to deliver bad Web Services. Whilst Web Services remove many of the technology constraints of communication between applications providing flexibility at the implementation layer, the business agility that is promised is more a factor of Service design than protocol adoption. As such SOA should be thought of not just as a way of designing and documenting an "Architecture of Services", showing their relationships, dependencies, etc., but also a discipline by which we ensure that those Services are the right Services, delivered at appropriate levels of granularity, abstraction and generality that makes sense to both Service Provider and Service Consumer, reduces the effort (particularly on the consumer) to use a set of services to perform a particular objective, and truly minimises the impact of change allowing Service consumers to switch providers and Service providers to switch implementations. We believe that SOA disciplines will become vital in delivering external Web Services where agility and flexibility is required (by both provider and consumer), and for Enterprise Wide Web Services where broad applicability must be ensured to enable reuse.
The Costs of Web Services and SOAMany of the benefits outlined above imply reduce IT costs resulting from the adoption of Web Services. What, if any are the additional costs of using Web Services and adopting SOA? Thanks to the universal adoption of Web Services by vendors, much of the software infrastructure required effectively comes as part of the regular upgrade cycle of existing products that most organizations will already have. This is not to say that organizations will not take the opportunity to decide whether they need to change products or vendors in order to obtain software better optimized for Web Services, or acquire additional products that might make them more productive. For example, some organizations might introduce new capabilities they have not used before such as business process orchestration/workflow products that now support Web Service protocols. However, the bottom line is that most of the essential Web Services capability can be acquired via software upgrades that also contain other useful functionality, and as such comes at effectively zero or little additional cost in terms of software.
The Cost of SOADelivery of the Web Services themselves in terms of the necessary protocols if largely hidden from developers by tools, and is unlikely to be an overhead in terms of programmer productivity. As such, much of the development overhead of delivering Web Services will come in the architecture, infrastructure, analysis and design phase to ensure that where required Services are
Performance OverheadsWhilst there are ways of optimising the use of XML, many Web Service scenarios will involve extra process steps that will likely add some performance overhead to the overall process. That said, existing alternative integration options such as EAI and B2B have similar performance profiles in comparison to a point to point connection between two applications using the same (proprietary) technology. Some possible sources of performance overheads will be,
Of course these should all be seen from the perspective of the benefit they bring, not just the overheads they incur.
NOTE: This report updates the original draft published with the title: Inside Every Web Service is a Benefit Struggling to Get Out
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
![]() |
© Everware-CBDI Inc 1999-2009 |