Talk about shooting fish in a barrel.

Shane K Johnson, Technical Marketing Manager at Red Hat wrote a fun little blog post comparing JBoss and IBM WebSphere in terms of total acquisition costs. Of course, it becomes a pretty unfair fight when you sit there comparing an open source product with a proprietary product designed to generate mounds of revenue from clients, but such comparisons are amusing nonetheless, and more importantly, they should get us thinking about exactly what it is that we’re getting from proprietary applications servers: a question that this humble editor still has a hard time answering.

The verdict? The WebSphere Application Network Deployment edition supporting 128 cores would run you about $2,489,000. Running on 256 cores, you'd be looking at a price tag of $5,893,800. And for JBoss EAP? You're only looking at $108,492 and $216,984 respectively. Now that's some serious cost savings.

Of course, IBM insists that the total cost of ownership (TCO) for an IBM product will be much smaller over time. Personally, I've always found these TCO arguments to be bunk. How quickly do you rack up an extra two million dollars in costs using JBoss vs. WebSphere? Heck, just the interest on that money could hire you a full time admin to babysit the JBoss server if you were really that worried about it.

Open Source no longer means Open Scare. With the delta between open source alternatives so large, it's hard to understand why anyone starting anew would even think about embracing a proprietary application server.

Head to Head: IBM WebSphere ND vs. JBoss EAP

I've used WAS 3, 4, 5, 6 and 7. If you factor in the time spent (ie wasted) waiting for WAS to start and deploy, the cost of WAS is atleast 3x greater than JBoss, Glassfish, and Weblogic. If you factor in the cost of re-installing WAS every few months on development machines, you can increase the cost.

But honestly, why bother comparing yourself to WAS. Please aim higher. Comparing yourself to the most painful java application server is like comparing yourself to hobbling. Things that normally work on Glassfish, JBoss and weblogic always end up needing tweaks and changes. Even though I make a living supporting and using WAS, honestly it is not enjoyable.

Aim higher please, don't compare your product WAS.

Is it 1912 ?

Are we still trying to sell application servers ?

Like kicking in open doors. Just the extra time needed to develop anything on WAS compared to JBoss or Glassfish outweights the total cost of JBoss EAP :-)

Instead of rah-rahing the post, it should be looked at the same way most vendor's claims are.  The basis of the claims are based on TCO studies that are slanted: IBM's are biased because the testers were ignorant about JBoss, and RedHat's are biased because the testers were ignorant about WebSphere (despite claims in each study to the contrary, it doesn't take much to see the b.s. in both directions).

I am no lover of IBM, but I cannot never figure out why so many people take a one-sided view of this question. Do any of you think that a company would come to a decision to choose IBM Websphere, and spend all that money, without some core fundamental reasons. I wish I had a dollar for every post about how JBoss or this app server or that app server is so much better and easier than Websphere. They opine about how it is such a non comparison - exactly like this article does. However, Websphere just had another spectacular quarter, racking up huge gains for IBM in revenue and profit. This was like the 10th year of stupendous growth... so I ask the question. Why is that? They just have money to burn? Their people are just stupid? Or is it that there are some fundamental reasons that companies are willing to spend that time, money and energy?  The market determines this stuff - and it makes decisions for a reason.... As a capitalist, I believe in the overall decision making ability of the capitalistic system - do all of you know something the system doesn't?

As an Architect, I worked in all major servers and it has its own limitation and advantages… BUT I would say ROI would be a critical factor which deeply looking into it… As a startup company, people would like to spend less money and gone with open source platform but they would have hired dedicated admin and support team for maintenance for these servers… If company grows then they would need to address large audience and stake holders for scalability and extendibility of the applications and further road map so they adapt single partner who does hardware, hosting and software’s... So they willing to spent money to satisfy their stack holders…but they may think that they will get their ROI by gaining confidence… BUT as a technical guy, I would opt for open source which would work with reasonable limitation which may take care by guys like US anyways….

This is an especially critical point.  IBM's marketing has moved up the value chain.  Instead of selling an app server, IBM sells systems and packages.  Today they announced a new line of integrated systems that build on their pattern-based deployment offerings, that reduce the amount of time and resources needed to deploy and care for infrastructure.

This is an especially critical point.  IBM's marketing has moved up the value chain.  Instead of selling an app server, IBM sells systems and packages.  Today they announced a new line of integrated systems that build on their pattern-based deployment offerings, that reduce the amount of time and resources needed to deploy and care for infrastructure.

That sounds good allright, "pattern-based deployment offerings". Problem is, in my experience, that this is just IBM-speak for poorly stiched together components, half of them acquired from other developers and then shoe-horned into some "product package" or "offering". Deployment and administration is a nightmare as tasks like deployment and configuration are sometimes done with different patterns and tools even for modules within the same component.

The WebSphere Portal/WCM stack (or whatever IBM calls it's "collaboration" offering these days) is a prime example of a bloated mish-mash of different technologies, groaning under it's own weight of technical and architectural debt.

I am no lover of IBM, but I cannot never figure out why so many people take a one-sided view of this question. Do any of you think that a company would come to a decision to choose IBM Websphere, and spend all that money, without some core fundamental reasons. I wish I had a dollar for every post about how JBoss or this app server or that app server is so much better and easier than Websphere. They opine about how it is such a non comparison - exactly like this article does. However, Websphere just had another spectacular quarter, racking up huge gains for IBM in revenue and profit. This was like the 10th year of stupendous growth... so I ask the question. Why is that? They just have money to burn? Their people are just stupid? Or is it that there are some fundamental reasons that companies are willing to spend that time, money and energy?  The market determines this stuff - and it makes decisions for a reason.... As a capitalist, I believe in the overall decision making ability of the capitalistic system - do all of you know something the system doesn't?

I've used multiple versions of weblogic, jboss, glassfish and websphere over the last 10 years. Each container has quirks and weaknesses. Having built and supported container agnostic applications, my bias opinion is that Websphere is by far the most painful EJB container on the market. I can deal with bugs in glassfish, geronimo, tomcat, and jboss, since I can debug it myself. I've found bugs in older versions of Weblogic and other containers.

My opinion is based on real world experience, not third hand stories. Even though I make a good living supporting and developing software on WAS, it is far from a "joyful" experience.

This would surely be an amazing and amusing comparisons.

FXDD