| |
Thursday 22nd May 2003
COMMENTARY - THE LONG AND WINDING ROAD
| CONCLUSIONS
So don't be surprised when you hear the major vendors de-emphasizing Web Services. They know they have over reached themselves on this issue, and that they need to redirect attention for a time. On demand, demand computing, efficiency and cost reduction are all good themes that play to today's economic circumstances. This is where the vendors will address their efforts, and not unreasonably establish the necessary operating environment for a world where use of services is pervasive.
For the same reasons the vendors are characterizing Web Services as "just another form of middleware". They realize they have to adjust their messages for a while, and play down the strategic importance of Web Services, until they are ready for mainstream deployment. So focusing on just another form of middleware is resetting expectations to be relevant to applications that can reasonably be implemented during our Phase One and Two.
Of course our readers will understand that this is a longer term game we are playing. Web Services is merely a phase in a longer running process which will eventually deliver a comprehensive service oriented environment. It's going to take time, but there's much that can be cost justified today, which equally contributes building blocks to the overall project.
Related Report
ROADMAP REPORT - A WEB SERVICES MATURITY MODEL
----------------------------------------
Adopting Web Services and SOA is an evolutionary process for vendors and their customers. The question is how do you manage progress and risk in a constantly changing environment? While the technology issues are inevitably dominant today, this is merely a symptom of the immaturity of the service environment. Very soon business issues will dominate. In this report we introduce a high level maturity model that provides a communication vehicle for aligning business and technology roadmaps. Report available to Silver and Gold members at:
CBDi Report | Abstract: Have you noticed the vendors starting to play down Web Services? It's becoming clear that the overall vision is further away than anyone realized. So what's reality? We provide a roadmap to assist your planning.
The course of history has never been straight; it only looks that way with the benefit of hindsight. The evolution of the archetypal IT environment has been littered with failures and false promises, and we only have to look at the layers of product, platforms and applications to remind ourselves that today's legacy is yesterdays silver bullet. Perhaps we haven't quite reached that point yet, but very soon the role of IT archaeologist will be commonplace, digging through the layers of technology in an attempt to understand the practices of earlier enterprises.
At this current point in time we have an interesting long running process that unusually provides us with a good indication of what capabilities will become mainstream in the years ahead. That long running process is the various standards development initiatives focused on WS protocols as well as grid computing and operational environments. The wide industry support being given to most of these processes, combined with a gradually shrinking number of key industry players, enables us to make forecasts with greater precision than we have hitherto been accustomed. It's useful here to consider a number of significant phases.
FIRST PHASE
The WSnn protocol efforts in the W3C and more recently in OASIS have now been running for some 3 years, and the all important trio - SOAP, WSDL and UDDI, together with the important WS-I profiles allow a basic level of description and interoperability of messages that establish a first base. We might regard this as completion of a major phase of work, that will permit certain types of application, which might be more easily defined by what they do not permit - for example reliability, transactionality, security etc.
In this phase the predominant service deployments will be:
- mostly internal
- using existing security mechanisms
- not mission critical
- focused on better application integration
CBDI has reported on numerous case studies that have shown how Web Services can be used to good effect with the basic protocols. For example we might instance Amazon that is very clearly set on reengineering their business and enabling their affiliates to do the same. But these are in a minority. Whilst there will undoubtedly be some new and innovative uses of Web Services, this phase will be best characterized as a period in which many enterprises do their early learning, and some existing applications are wrapped in order to establish some elementary level of service oriented architecture.
It's clear there's a big difference between adopting Web Services and becoming a service oriented organization. Web Services in the end are simply a better form of middleware. Moving beyond that level requires a change in business practices not just technology.
SECOND PHASE
It has become something of a "cause celebre" that the absence of Web Services security has been a significant inhibitor to adoption. However this bears re-examination. Those that have wanted to secure Web Services using what we have defined as first phase protocols, have used pre-existing security mechanisms (SSL, CORBA etc) without difficulty. What the WS-Security protocols bring is heterogeneous message level security, which we suggest is relevant to more sophisticated architectures that also need enhanced guarantees of reliability and availability.
And it seems most likely that the combination of security and reliability will form the basis for a second phase of Web Services adoption that will be characterized by the ability to offer some guarantees around service levels. It also seems likely that this phase will be characterized by some level of business process integration. From a timing perspective it seems highly likely that the BPEL specification will prevail and will, by the early 2004 timeframe, provide a reasonable basis for business process integration. This is corroborated by the primary vendors supporting BPEL that plan to have tools available in this period.
In this phase the predominant service deployments will be:
- business process oriented
- project level implementation
- mostly internal usage
- based on a more mature understanding of SOA with better separation of layers as BPEL scripting is implemented
THIRD PHASE
For some time now there has been evident something of a disconnect between the ambitions of the major platform vendors and the standards processes. The issue here is management, the protocols which are being worked under the OASIS WSDM committee.
The issue is that enterprise level services will only become acceptable to enterprises when the provider can provide rock solid guarantees of availability, reliability and performance, AND can dynamically manage the environment to deliver that, AND prove in retrospect that the commitments have been met. This will be the major inhibitor to more widespread use and reuse of services. Although there are an extraordinary number of ISV's that have entered this Web Service Management marketplace, the typical enterprise is going to a) see a standards based platform as a prerequisite for implementation and b) prefer to wait and see what the major platform vendors provide. Early adopters will always move with early adopting tools vendors, but the mainstream market will wait.
The current emphasis placed upon on-demand or demand based environments is going to be a critical enabler of the end point service vision. What's interesting is to compare the efforts of the major vendors. In our previous reports on this area, we have characterized the efforts of HP and Sun as primarily about utility computing - creating a flexible technical environment, whereas IBM is increasingly focused on the business opportunities that are enabled by on-demand environments. This increased visibility is going to be essential in our Third Phase, as we have pointed out in the recent CBDI Newswire - Groundhog Day. As one correspondent accurately commented, it's going to be essential to "Convince the main board to address enterprise IS architecture seriously, not as a modeling exercise and not as a means of controlling IT but as an integrated part of business strategy planning."
Our Third Phase is therefore the point in time when service oriented applications are fully integrated with the business from organizational, funding, and product development perspectives, to name just a few. In this phase the predominant service deployments will be:
- enterprise level, with common services used right across organizations
- services implemented as an integral part of business products
- supported by guarantees and standards based measurement and monitoring systems
- enabled by a wider selection of available services both inside and external to organizations
OMMISSIONS AND TIMESCALES
The IT archaeologist of the future might be a little puzzled when he or she sees a huge amount of hype and spin about Web Services between 2000 and 2003, which then dies down considerably. What's happening here is that the industry actually didn't over hype the "potential" for Web Services, but what they "did" do, was to set the expectation that the vision would be delivered in a relatively short timescale. The reality is very different. We are currently in Phase One as we have defined it. Many organizations will only enter that phase this year and next. Phase Two will commence in 2004 and run through 2007. Phase 3 will commence sometime perhaps late 2005, and run and run.
It is also important to recognize that the standards processes as we see them today are not the end of the line. As with every standards effort, these are going to continue. One of the most important continuing efforts, and conversely the weakest of the current standards efforts, is the whole area of process choreography. Whilst BPEL is considerably better defined than the competing specification WSCI, it does not deliver a comprehensive view of the contract between the provider and consumer.
The importance of contracts should not be underestimated. It is very difficult to adopt an on demand approach (to either computing resources, or business processes) without having a very clear and precise way of expressing machine readable contracts. This is a major failing, because without it, there is going to have to be incremental communications between the collaborating parties. Widespread reuse of services will only occur when the consumer is given a comprehensive contractual view that documents all the behaviors and obligations, in other words a specification model of the externalized perspective of the offered service. This is perhaps the next big challenge, and as yet this requirement is not yet even comprehended by some major players in the modeling tools space. |
| | Read Full Article / Link to referenced materials |
|
Home