Discussions

News: IDC Report: Web services a decade away

  1. IDC Report: Web services a decade away (14 messages)

    Web services are proving their value today when deployed by early adopters to integrate heterogeneous systems in highly decentralized organizations with multiple locations. However, the next phase for Web services ? delivering software as services ? is at least a decade away, according to IDC.

    Read Report: Web services a decade away.

    Read IDC Offers A Dose of Reality To Remedy Overly Optimistic Web Services Hype.

    Threaded Messages (14)

  2. What are they drinking?

    IDC,Gartner, et al, these analytical organisations are allways making predictions, often predictions which seem to me ( with 10+ years in the industry ) to be unrealistic ( read stupid ).

    Has anyone ever done a study on how accurate these organisations are? According to most the Y2K problem was supposed to return us to the dark ages :-)

  3. Yeah. One day it's "next year" the next it's "ten years away." Before you know it - it's already happenned.

    (That must be a Yogi Berraism or something).

    Regards,

    Rich
  4. IDC Report: Web services a decade away[ Go to top ]

    I have just implemented Web Services for a Commercial Insurance system, which are being used by third party software systems to transact business.

    I was pleased to hear that it's 10 years ahead ! :-)

    Anyone else implemented commercial web services recently ?
  5. IDC Report: Web services a decade away[ Go to top ]

    Yep. We do it to a certain extent now externally. One project that I've just left used webservices internally for all business logic calls (although the thing was designed and implemented before SOAP existed, really). I'm astonished that anyone could think this technology is 10 years away given it's usable /now/. Once we have 2PC working, it could get /really/ interesting.
  6. IDC Report: Web services a decade away[ Go to top ]

    One project that I've just left used webservices internally for all business logic calls


    Why do that? It should be done as needed not as the default. If the call is local, it shouldn't be done via web services.

    They did say it could be used now. They said the full vision is years away.
  7. IDC Report: Web services a decade away[ Go to top ]

    I can see this is going to be another battleground issue: Performance cost vs. flexibility of webservices.

    Why do folks fight evolution tooth and nail? we saw it with all of these:
    Dos vs. Windows
    C vs. assembler
    Java vs. C++
    J2EE vs. .Net
    EJB vs. local classes.

    Objectively, the more you optimize for performance the less dynamic behavior is supported... such as garbage collection, portability, remotability, security
    enforcement. At some point even basic features such as reflection and inheritence impact performance...

    Inevitably, Moore's law along with murphy's *always* causes flexibility/adaptability to win-out over performance.

    Take Biological evolution for example:
    (Humans take longer to become adults than other animals but the delay is made up for with the enhanced ability to learn and adapt...same for highly
    evolved technology.)

    Matt
  8. IDC Report: Web services a decade away[ Go to top ]

    I can see this is going to be another battleground issue: >Performance cost vs. flexibility of webservices.


    I don't see webservices as more flexible. While the solve some problems they create more.

    >Why do folks fight evolution tooth and nail

    Currently, webservices is sort of a step backwards. We are back to passing data and not objects.

    Web Services have their use. And they should be used wisely. And I am not talking about peformance but flexibility and realiability
  9. IDC Report: Web services a decade away[ Go to top ]

    I would be interested to hear how you and the third parties are handling exceptions, invalid objects/data and general OO principles (i.e. using interfaces). I assume you are using Java. Are the third parties using Java?

    Not that I agree with these prognosticators, but you probably haven't implemented the full vision of web services (I don't think the web service vendors have yet). Not that what you have done isn't pretty cool.
  10. IDC Report: Web services a decade away[ Go to top ]

    <quote>
    I would be interested to hear how you and the third parties are handling exceptions, invalid objects/data and general OO principles (i.e. using interfaces). I assume you are using Java. Are the third parties using Java?
    </quote>

    Yes , it's Java (and SOAP over HTTP) ! Exceptions are handled by simply returning a SOAP Fault with a predefined message. The messages are relatively complex (being commercial insurance) so we validate each message against an XML schema, this avoids 95% of invalid object / data problem. The other 5% validation is coded. We have used design patterns and good OO principles as much as possible in building the service, but of course this is hidden from the client systems. The two primary third party systems currently using the service are built using MS VB.

    <quote>
    you probably haven't implemented the full vision of web services (I don't think the web service vendors have yet)
    </quote>

    We have constructed the Service ourselves rather than using a product supplied by a web service vendor. There are, as always, improvements that can be made, but the service is now trading.

    Regards,
    Greg.
  11. IDC Report: Web services a decade away[ Go to top ]

    The IDC report seemed to be saying that the "high water mark" of Web Services is 10 years away. To me, that's different from saying, "Web Services are 10 years away." I interpret as saying that the zenith of the technology is 10 years away, after which it will slowly be replaced by something else.
  12. All I would say is look at the last 10 years or so and see how much things have evolved and changed. The volatile nature of this industry suggest that one will be very lucky to even predict accurately the future of a technology for next couple of years. 10 years is way out of sight.

    I guess the point they were trying to make is that there is still considerable amount of work to be done in this area.



  13. I don't know where these organizations come out with some vague ideas like this. All the factors they take to give such an opinion is based on the current technology. What if there's one more revolution like Java, which could totally change everybody's beliefs? You can't rule that out. And, that time, all these reports go for trash. I feel these IDC or Gartner should stick to predictions not more than a year.
  14. I think it's important to note the difference between the technology that Web Services are being built on, and the original vision of Web Services itself. As I see it, one of the major goals of Web Services is to enable dynamic lookups and retrievals of the necessary interfaces to execute code remotely. This is the part that is stil not very practical, I think. One of the reasons for this, is that I think that this capability is a solution looking for a problem. Once people begin to have a real need for this in their applications, the implementation will arrive much sooner than 10 years.

    When people say that they are using Web Services today, I think that this typically means that they are using SOAP, WSDL, etc. Basically it's just a convenient way to execute RPC. I can name a number of applications and companies that are doing this already. But I don't think it's fair to call them full fledged web services. XML-RPC is a much more appropriate label for what is happening now, IMHO.

    Ray
  15. In my view, the comment "I think it's important to note the difference between the technology that Web Services are being built on, and the original vision of Web Services itself" is spot on; there's a lot of SOAP in production out there, but the vision of Web Services is not RPC.

    If you look at the content of the recent additions to the Web Services specs (transactions, process modelling etc.), they present a vision that would standardise application integration and business process modelling across multiple platforms and middleware. This is a long way from the current (and very fragmented) market of integration technologies, particularly those based around process modelling (this isn't just J2EE and .Net). The two things you gain from this that start to back up the Web Services promise are a). interoperability and interchangeability between whatever integration solutions you've built in the past (or acquired in mergers etc.), and b). the whole lot will work over the existing network of routers, workload managers, failover switches etc. to handle HTTP traffic, JMS traffic, e-mail traffic etc. Add those together with the appropriate process (re)engineering, and you'll have a much more flexible IT and application infrastructure supporting your business than you do today.

    So, when would you expect a good proportion of companies to have integrated the majority of their appropriate internal and external interfaces like that? A decade sounds as good a bet as any to me. In the meantime, all the individual implementations to get there will keep us all plenty busy ;-)

    Rick