Thoughts on Containerization of the Webmethods

Thoughts on Containerization of the Webmethods

I was lying in bed the other night thinking about Crossvale’s history.  On our first day of business we met with a little integration company (back then) called webMethods.Service oriented architecture (SOA) was hot.  Business process management (BPM) was gaining traction and face to face classroom training was how education was done.  We came out of that meeting with a handshake on a deal and we were off to the races!  {This is long before SAG bought webMethods.  I was young, naive and ready to roll.  In truth, looking back, it was total madness! ????  }

Anyway, the better part of two decades later and here we are embracing the latest wave of technologies.It is a “perfect storm” .  Open Source is now regarded as the way to buy solid, supported software.  The cloud is gaining more adoption every day and containers are an unstoppable wave..Red Hat OpenShift being the container platform of choice means that (thankfully) Crossvale’s services are in high demand.   We are, after all, an APEX partner and winner of the Red Hat Application Partner of the year.However what got me thinking was “what about all those guys who have big investment in SOA?  What about the webMethods guys?”Thankfully they have a solution – Containerize the webMethods Integration Server.Some of the steps are highlighted in this web page.The wM Integration Server (IS)  is fat.  There is not much dispute about this. The IS is well designed for a monolithic SOA and in truth Crossvale had a lot of success with the wM IS when fat monolithic enterprise service busses (ESBs)were the newest and greatest technology.  It is still employed a lot to day and is well received.However, I digress…  Dockerizing the fat IS is somewhat contrary to the latest architectures which lend itself much more to micro-services (before any of you comment, a fat wM Integration Server with one service deployed on it is not a micro-service).So why are folks doing it?  Investment, spend and dependency are the obvious answers.  An enterprise could be totally dependent on the wM platform, have spent millions building out a service tier but want to move to containers for all that they offer.  Throw away and start again with micro-services on, say, Spring Boot is just not happening.  It can’t.  The enterprise must continue to move forward.This is not to say that a containerized webMethods IS sitting along side spring boot micro-services could not be employed.  It absolutely could and would work well.Anyway, it is good to see that the webMethods IS can be containerized and is employing this latest docker technology.  If I was handling it I would for sure drop it on Red Hat OpenShift as opposed to a more basic Kubernetes cluster.  The value prop of OpenShift is well documented.If you want a hand getting webMethods containerized with best practices employed, or just want to talk webMethods, docker, kubernetes and Openshift in a single conversation ( or single sentence) drop me a line.  I’m always glad to kick a few ideas around.In the mean time I think I’ll lie back this evening and think of my “good ol’ days” when I was cutting code for a living.  SOAP services, Flow programming language, basic auth and all the rest of the goodies…………………

Conor Brankin, 

CEO, Crossvale Inc.