Discussions

News: What is the state of SOA? Take SearchSOA.com survey

  1. Service oriented architecture (SOA) has had some ups and downs over the years, and even quite recently. Certainly, developers, architects and IT can have divergent views of 'what it is.' We'd like to know what you think. TheServerSide.com's sister site SearchSOA.com is conducting a survey on SOA, and would like you to take part. Respondents are eligible to enter a drawing to win a $100 Amazon.com gift certificate. Individual responses are confidential. To take part in the survey, click on the link below. Thanks for helping us. http://www.surveygizmo.com/s/99337/tss-soa-survey-2009 Meanwhile, do you have comments on the SOA phenomenon? On one level, 'services' are not terribly different than components or object technologies of the '90s. Yet, several key Java vendors have near-fully endorsed the SOA mantra.

    Threaded Messages (12)

  2. Just like object-oriented and component-based programming in the 90's, SOA has been oversold as a silver bullet. First, SOA is an *architecture*, not a technology. Vendors have been hawking their products under the SOA banner by trying to convince companies that it'll solve all their problems in a minute. The truth is that SOA is hard. The technology part is fairly straightforward, although thinking in terms of loosely coupled components is difficult for many. However, it's the organization that's often the major hurdle. Many enterprises have been building stovepipe systems for so long, and the organizational walls that go with it, that they just can't tear down the walls an collaborate. Services require a company to either organize around services, or agree to share the code and enhance the services based on functional need. In the first case, this requires cooperation and coordination between groups. Yea, right. Too many different agendas lead to a dead end. The 2nd approach (shared code) is often verboten, as there's no single neck to wring. Mark
  3. I think you cover essential points. Most of what is worth doing is hard work - yet software vendors have a vested interst in making it seem like magic [not work]. And doing services consistently overtime? It depends on how people work together or dont - in other words it depends on 'the organization.' "Peopleware" is a timeless book that speaks to this.
  4. State of SOA = State of the Union[ Go to top ]

    Heavily oversold, In-debt. Companies who didnt know what it really is and meant, started using it as a SilverBullet and so have seen SOA-at-their-place turn into a Sink, sucking in all cash-resources.
  5. Just like object-oriented ... programming in the 90's, SOA has been oversold as a silver bullet.
    Object orientation has never been oversold, and still is a silver bullet.
  6. Object orientation has never been oversold, and still is a silver bullet.
    Considering how non-intuitive OOD is, I totally disagree. Like SOA, it doesn't really buy you much, and you can just as easily write unmaintainable apps in OOD as procedural languages. Keep in mind that many COBOL and other proc apps still exist. OOD was designed to aid with issues like encapsulation and access issues, but a silver bullet it is not. Look at the reasoning behind AOP
  7. Object orientation has never been oversold, and still is a silver bullet.


    Considering how non-intuitive OOD is, I totally disagree. Like SOA, it doesn't really buy you much, and you can just as easily write unmaintainable apps in OOD as procedural languages. Keep in mind that many COBOL and other proc apps still exist.

    OOD was designed to aid with issues like encapsulation and access issues, but a silver bullet it is not. Look at the reasoning behind AOP
    Do you still develop software in the procedural style?
  8. Go with the flow[ Go to top ]

    Do you still develop software in the procedural style?
    No, but is your code bug free? OOD is not a silver bullet, and to be honest, none of the platforms and frameworks are. If they were, Java IDEs wouldn't come with debuggers. Back to your point, I'm not paid to do it procedurally, but I have friends that most certainly are in the realm of scripting, COBOL, JCL, and the list goes on. Also back to the point, since refactoring tools have become mainstream, I write more procedural code now than before because I can just refactor it to something more logical if needed. Proceedural is intuitive and OOD isn't. OOD is more reuseable, easier to structure, and more maintainable though. Neither are silver bullets. What ever happened to CASE tools anyway? I want my laptop to write code for me. :)
  9. Re: Go with the flow[ Go to top ]

    Do you still develop software in the procedural style?

    No, but is your code bug free? OOD is not a silver bullet, and to be honest, none of the platforms and frameworks are. If they were, Java IDEs wouldn't come with debuggers.
    silver bullet != bug free; OOD == object oriented design != object orientation; debuggers == fun in an object orientation environment;
    Back to your point, I'm not paid to do it procedurally, but I have friends that most certainly are in the realm of scripting, COBOL, JCL, and the list goes on.
    scripting == ruby == object oriented; Java == the next COBOL; JCL == ???;
    Also back to the point, since refactoring tools have become mainstream, I write more procedural code now than before because I can just refactor it to something more logical if needed.
    more refactoring == more object orientation; more code == more practice == more object orientation;
    Proceedural is intuitive and OOD isn't. OOD is more reuseable, easier to structure, and more maintainable though. Neither are silver bullets.
    more reuseable, easier to structure, and more maintainable though = silver bullet;
    What ever happened to CASE tools anyway? I want my laptop to write code for me. :)
    CASE tools != object orientation; CASE tools != silver bullet;
  10. Re: Go with the flow[ Go to top ]

    OOD is more reuseable, easier to structure, and more maintainable though. Neither are silver bullets.

    more reuseable, easier to structure, and more maintainable though = silver bullet;
    My understanding is that 'silver bullet' in the context of software is more than just an improvement in certain areas. If I'm not mistaken, the term comes from a Brooks essay "No Silver Bullet". The point being that there was no tool that would eliminate the big problems in software development. So most people take 'silver bullet' to roughly mean 'panacea'.
  11. Is SOA hard?[ Go to top ]

    There is no answer to it..rather this is not the logical question. beacuse neither there is a tool to check if the deployment of SOA "enabled" product/architecture is right or if there is any compliance authority to measure that. Various SOA evanglists have proposed various models/SDLC for SOA but sometime I think do they really know what they are talking about. I dont know either..but I have seen people mixing "web services" with "services" with SOA. SOA is not hard or soft.Its the concept of re- architecturing, brainstroming and finding out how to reuse the existing system to its fullest...may be more than that... ~Lokesh Pant http://lokeshpant.blogspot.com
  12. The benefits of an SOA environment can transform how a company does business and how IT supports the business. I have been in the IT industry for 20 years and I have not seen a transformational technology approach as large as SOA. The move from structural programming to OO is close. The deal is, if your company is poised with a set of problems that an SOA architecture can help resolve, SOA a is a "Golden Bullet". An example of a Usage Pattern is an organization with tightly coupled applications in an industry of great change. This pattern is amplified if the number of applications that IT supports is large, such as thousands of applications. There are many reasons for failure and in-efficiencies is implementing an SOA infrastructure. Very few of these revolve around the technology tools. People (organization, education, and experience), governance(developmental and operational), and lack of documentation are some of the main factors to failure. Experience is worth it's weight in GOLD GOLD GOLD for success. Also a plan! Every company is different. My current company implemented different portions of the SOA stack over 18 months to assimilate the organization to the paradigm. [ Orchestration, BPM, Complex Event Processing, Composition, Mediation, than Virtualization) while implementing governance initiatives along the way. At the end, Agility and cost reductions are being driven by Enterprise Architecture by Usage Patterns and Reference Implementations with our selected technology stack of choice, to solve the businesses problems. Compared to sharing Database connection information across a global enterprise and making our applications logic and data process spin out of control in a spider web, we have a unified approach to drive success. GOLD GOLD GOLD. Good Architects and an understanding of the "Current State" and vision of the "Future State" and an architectural plan of how to get there has been the key for any initiative.
  13. wth[ Go to top ]

    Seriously... I come here for news, not to read spams about surveys to win prizes. Beyond that, SOA is an overly used approach by 90% of the companies so they can be "SOA" oriented. My experience has been that SOA yields little value and causes far more headaches and lost dollars due to people that claim to be architects having no idea what they are doing. If your enterprise architect doesn't have an SLA with the teams implementing his/her vision, then you are wrong and headed for pain.