Welcome to The Metaverse : Service Orientation

WCF Magic8Ball Sample Updated

richardt — Wed, 30 May 2007 05:05:00 GMT

I realized today that the WCF Magic8Ball sample app that I posted some time ago is now woefully out of date, so I spruced it up a little and posted an updated version of the source.

Enjoy!

Where do I stand today on ESB and the mythical "successful big project"?

richardt — Fri, 11 May 2007 03:22:00 GMT

I received an email from a long-time follower of my blog that I thought might serve to spark a little interesting debate and also serve as a marker in time for my opinions on the subject of ESB and the ever elusive "successfully delivered big project":

Hi Rich-
Followed your blog from loosely coupled. I am reading these thoughts about the ESB from your post and the loosely coupled post, right around the same time, I had been pushing the same thoughts at various Government of <country> project that I was involved with....somehow "figure stuff out first, build some stuff, and then worry about all this big infrastructure stuff" just does not appeal to people. Hence I have seen two rather large ESB projects crumble under the weight of the huge initial investment and the lack of interest from would-be users.
So I am interested in finding out....2 years after your "debate" with Dave Chapell [part1, part 2], where do you think the ESB situation is? Are the WS-* standards that mature yet? I am once again on a similar path, preaching the same message...
Would love to hear where you are with the ESB stuff now?
--cheers
<name>
<name>@<domain>.com
SOA Architect

To be honest, I don't see any change in my perception on this matter because my beliefs back then were based on painful experience. I haven't yet seen any revolutionary improvement in the way projects are executed that leads me to believe that someone has thought up a significantly "better way". This is largely because projects are lead by wetware ... wetware with egos, grudges, issues and which can be easily distracted, diverted, disillusioned, etc.

And so, as you've pointed out, we continue to see repeated implosions where projects to boil the ocean with a shiny new candle collapse under their own gravity. I don't know when people will start to wake up and smell the White Chocolate Latte (yum!:)) and realize that taking smaller, more achievable, more manageable steps which allow you to course-correct while you proceed, whilst also having to accept that you're going to learn stuff along the way which may lead you backwards on occasion ... is the only way we're going to proceed.

I said it before and I'll say it again - mistakes, learning and progress are not mutually exclusive - they're joined at the hip. There is absolutely no way you an architect something today that takes more than 10 months to build and expect that you will have designed and built this thing to withstand every scenario you want to throw at it - by the time you're done, the world will have changed considerably.

This, to me, if the crux of the issue - Stuff changes ... deal with it! Tools change, platforms change, ideas change, markets change ... and they all change faster now than ever before. Assume the velocity of change is only going to increase and you'll be better prepared to handle real life! :) If you don't and continue in the mistaken belief that the world will be the same 12 months from now, you're asking for trouble.

To the question on whether WS-* is mature, I'd definitely say that the WS-* protocols are significantly more mature than they were 2 years. Many are now onto their 2nd or 3rd revision after being used in production systems and products for 3-4 years. Several more have been through the wringer over the last few years and are now being submitted to standards organizations now that we have a pretty good feeling that they've reached an appreciable level of maturity.

Remember also that many of the WS-* protocols were born from a synthesis of what experience has taught us were already effective ways to get certain things done, so it wasn't like they were just created from an "ah-hah"! While all the usual discussions are still to be found on the forums and on the blogs. E.G. "Is WS-* as efficient on the wire compared to binary protocols", the usual answers still apply - the size of the data on the wire should rarely be a true inhibitor to adopting these protocols. This sounds like an interesting corollary to my suggestion above that things change - and are changing faster: but recognize that some things (like human behavior ... the kind of behavior that commissions and executes projects) change at a glacial pace where any increase in the rate of change takes an eon to see.

So to close (and because I'm taking my wife out to the movies tonight and there's a Starbucks on the way with a White Chocolate Latte with my name on it), I guess I'd say that little has really changed in the guidance relative to this space in the last two-three years. Products and platforms have improved enormously, tools are more plentiful and more productive, techniques are honed, but only marginally so.

But people will still believe (or be convinced by others) that they have to go boil the ocean with a candle ... just be sure not to get wet - you might catch a nasty chill! :)

WS-Trust 1.3 and WS-SecureConversation 1.3 now standardized

richardt — Fri, 06 Apr 2007 19:42:17 GMT

There's a lot happening in the web-services standards world right now. A good example of this is that both WS-Trust 1.3 and WS-SecureConversation 1.3 have now been ratified by OASIS.

WS-Trust provides a simple protocol to allow someone to request a security token containing some set of claims from an Identity Provider (IdP). Because WS-Trust is part of the WS-* suite of protocols, it can be composed with other protocols to, for example, enjoy data integrity (signing) and data privacy (encryption) using WS-Security and WS-SecureConversation.

WS-SecureConversation adds to the capabilities of WS-Security and essentially enables the construction of a secure context that optimizes multiple calls between two parties.

These protocols are both used by Windows CardSpace, along with several other current and emerging technologies that enable dynamic, user-initiated identity federation.

Why is the ratification of these two important protocols a significant event? Well, as Gartner puts it:

"The availability of these new standards means that Web services security has finally reached an acceptable maturity level. The issuance and dissemination of credentials between different trust domains via an STS can now be achieved using a syntax that is familiar to most developers."

It's great to see the rich fabric of the WS-* protocols now reaching a level of maturity and sophistication that enable solutions to previously costly/difficult/impossible problems. Identity federation is just one example that many people don't even know they need, but once they start enjoying its benefits, will wonder how they did without it.

Designing Service Oriented systems just got easier!

richardt — Mon, 31 Jul 2006 23:15:00 GMT

[Updated 2006-08-02 with correct links]

When discussing how to design and build Service Oriented systems, one of the most enduring questions I've heard is "how do I design a Service Oriented system?" The actual meaning behind this question ranges from requests for help on how to approach the process of "thinking Service-Oriented", but more often the question often boils down to "how do I plan, design, layout and operate a SO system?"

The answer often depended heavily on what methodologies the questioner uses to design systems today. While many of today's design tools and methodologies can be applied to solving this problem, they are often overly complex and require significant skill and the training and practice that results in that skill) to render the resulting design document with sufficient accuracy and clarity to enable an independent consumer of that document to implement the given system successfully.

The difficulty in rendering sufficiently accurate designs using today's largely textual) methods has given rise to several Domain Specific Languages – ways of describing a design in a language that is (largely) unambiguous and broadly reproducible.

Several years ago, Microsoft started down the path of designing such DSL's as part of the overall Dynamic Systems Initiative (DSI). The core goals of DSI are essentially help simplify the design and operation of today's complex distributed systems far more effectively than ever before. DSI aims to provide a suite of DSL's that help us specifically describe what a system should do, how it should be composed and structured, how parts of the system communicate and relate to one another and what policies govern the operation of such a system.

I am extremely excited to see the release of the Service Modeling Language (SML) specification. SML is the result of the combined work of Microsoft, IBM, Sun, BEA, BMC, Cisco, Intel, HP Dell and EMC and is destined to be used to "used to model complex IT services and systems, including their structure, constraints, policies, and best practices".

Now, of course, design tools will help render such DSL's in a comprehensible manner - converting pointy-brackets to diagrams that mere humans can understand and I for one can't wait to see support for SML and other related DSL's being built into powerful tools by Microsoft and others.

WCF Magic8Ball Sample

richardt — Sat, 17 Jun 2006 02:34:00 GMT

[Updated 5/29/2007 to make the sample work with RTM WCF bits. Note that the accompanying article is a little out of date, but still substantially relavent]

Over the last year or so, myself and several colleagues have presented Windows Communication Foundation (WCF - formerly "Indigo") at conferences including PDC, TechEd, VSLive and to many user groups and customer audiences around the world. Throughout our on-stage code explorations of WCF, one of the most popular demos was our coded-live-on-stage Magic8Ball service and app. We’ve received literally hundreds of requests for us to publish the source and so, now that WinFX Beta2 is out the door, I figured it was about time to post an up-to-date article on how to build a Magic8Ball service and app using WCF.

Here is a ZIP file containing the PDF whitepaper and the source for this demo. Enjoy! :)

How Granular Should My Services Be?

richardt — Tue, 11 Apr 2006 00:00:00 GMT

This is a) a very common question, and b) very hard to provide absolute prescriptive guidance on.

In short, this is where the “art” of architecture & design comes into play!

How granular you make your service largely depends on how you’ll use it. Always bear in mind, however, that you should avoid traversing the wire as much as possible. Therefore, if you design your service so that you have to call 10 actions in order to refresh a customer data page, then you’ll likely experience severe perf and scale issues. A better approach is to call one GetCustomerData(…) method once and build the customer data page from the returned data.

“Ah, but my service has to be used by many callers – some require a great deal of customer data, but many only need the name of the customer”. In which case, your service should expose BOTH fine-grained and large-grained actions:

[ServiceContract]
class Customer
{
  //  Only returns customer name.
  [OperationContract]
  string GetCustomerName(Uuid customerID)
  { /* implementation */ }
  
  //  Only returns customer address.
  [OperationContract]
  CustomerAddress GetCustomerAddress(Uuid customerID)
  { /* implementation */ }
  
  //  Large-grained customer record retrieval.
  [OperationContract]
  CustomerRecord GetCustomer(Uuid customerID)
  { /* implementation */ }
  
  //  Large-grained customer record update – note that this is one-way!!
  [OperationContract(IsoneWay=true)]
  Void UpdateCustomer(Uuid customerID, CustomerRecord customerRec)
  { /* implementation */ }
}

Note that when using newer technologies such as WCF (formerly “Indigo”), anything that returns a void could be marked as OneWay. This means that the sender doesn’t have to hang around and wait for the service to complete processing the action. This can be a massive scalability booster! :)

New Direction

RichTurner666 — Sat, 18 Feb 2006 02:22:00 GMT

Long-time followers of my blog will no doubt remember when I started this blog in Feb 2004. I began by asking the hot question of the day: "What is Service Orientation" and "Is .NET Remoting Dead?". Since then I have continued to provided considerable guidance on how to position Microsoft's current distributed systems assets relative to the new wave of technologies in WinFX, as well as discussing many meta-issues such as the performance characteristics of distributed systems technologies, and much general discussion around SO, SOA and ESB.

Today marks the beginning of a change in this blog - as evidenced by the title change!

Today, I am happy to announce a change in position: I am now Product Manager of "InfoCard".

Now, before you ask, I'll will explain "InfoCard" in several subsequent posts, but if you can't wait to find out, I suggest you go read through some of the whitepapers currently posted to the "InfoCard" portal. If you don't want to read, then go here, register (it's free) and watch BillG's keynote where you'll hear and see some of Microsoft's strategies for protecting you from abuse online and from a variety of online identity fraud (including a cool demo of "InfoCard" demonstrated by yours truly).

Does this mean I'm not going to comment on WCF, SO, SOA, ESB, etc any more? You'd be that lucky! ;) Of course I will, but I'll also be interleaving discussion involving what Identity means, why it's important, what happens if we, the entire industry, don't work together to fix the current malaise of identity related issues ... and what happens if we do.

As always, I welcome your thoughts, ideas, questions and flames.

Okay, onwards into the Metaverse!