member login

WebServices dot org

Todays Featured Content:

Service Oriented Virtualization

SOA and Virtualization are currently considered to be two separate disciplines, but they no longer need to be. SOA offers the enterprise the benefits of increased agility and cost efficiency in terms of application development, reuse, and making connections across heterogeneous applications and business partners

iTKO LISA Combines SOA Monitoring with Advanced Test Execution Capabilities

Native test interaction with leading system metrics dashboards and reporting environments provides improved control over performance and reliability.

For SOA, The Future of Quality is Federated

This paper will refer to government organizations as a case study on SOA Governance. However, architects and developers in the business computing arena can draw valuable lessons from the complex integration and quality challenges faced by federal agencies.

iTKO LISA 4 Release Revolutionizes SOA Quality with Virtualized Services and Business Process Testing Features

LISA's Evolution Mitigates IT Risk through SOA Testing, Integration Support and Policy Validation

iTKO, Inc., the leading provider of testing solutions for SOA (Service-Oriented Architecture) software, announced the availability of the new version of its flagship product suite, iTKO LISA 4 SOA Testing and Validation. LISA expands upon iTKO's delivery of the Three C's of testing - complete, collaborative and continuous - by adding key functionalities that mitigate the business risk of ever-increasing change and complexity in enterprise IT.

Featured Content provided by iTKO

Your Policy Isn't Big Enough

John Michelsen
8th Feb 07:

As we continue to mature the SOA Governance space, the Policy area appears to be the one that is the most immature.

Vendors that are focused on Policy have a big, big job ahead of them.In fact, we believe that the Registry/Repository area and the Testing area of governance will have a lot to do with policy definition and enforcement in addition to those tools that are called governance policy products.

The reason is because policy is a VBT (Very Big Thing).In fact, I think 3 years from now governance will really just be defined by policies and in so doing it will take all three aspects of governance to make policy work.

A simple example -- today’s Policy typically sounds like:

  • “Interoperability checks”

  • “ws-security verification”

  • “Hit testing on XML documents (for you bit twisters like me)”

Pretty granular stuff. That’s not a very big policy.

In fact, the business won’t want to mandate policy that has much to do with things in XML and what standards you’re using.Of course they want standards to get the reuse and interoperability, but their version of policy will be:

  • “I really need functional integrity on this particular transaction activity of the system”

  • “I really need customer credit checks to always be accurate to within 30 minutes of the customer’s activity.”

  • “I can’t allow stale data to be reported as customer’s current activity.”

All of these kinds of things – that’s Policy.

When the business talks about policy, they are not going to read standards definitions from OASIS.They are going to talk about what their business functions are and what kind of expectations they want to put on the systems that are implementing those policies.That is Policy – that is a Very Big Thing.

So, short and sweet, if you are talking to a vendor who has got a so-called complete SOA Governance Policy Management System -- I haven’t yet seen that thing.Their definition of Policy is way too small.

You’re going to have to sign off on the much bigger problem of Policy at the business level, which is what SOA should be all about.Remember, we are aligning IT to the business, but not if we force-feed a definition of Policy to business. Business will take on a "free-market definition" of SOA Policy , whatever the consortiums try to dictate.

When the definition of policy is business-aligned it’s going to require very rigorous, here’s the bad word, “TESTING” -- on a continuous basis of the functional integrity of the system as it is running.Of course, things like IT guidelines are good policy, Performance Expectations are good policy, Down Time and Notification and all of those other things, very good policy.But if it isn’t also heavily involved in the way the business wants to set up expected behaviors, it’s not big enough.

"

Reprinted from http://itko.blogspot.com/

"


Trackback URL for this post: http://www.webservices.org/trackback/id/82241

Comments