WebSphere

Subscribe to WebSphere: eMailAlertsEmail Alerts newslettersWeekly Newsletters
Get WebSphere: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


IBM WebSphere software products Authors: Yeshim Deniz, hyper filter, Timothée Bensimon, XebiaLabs Blog, Javier Paniza

Related Topics: Java EE Journal, ERP Journal on Ulitzer, WebSphere

J2EE Journal: Article

WSJ-IN Exclusive: IBM Responds to Latest Microsoft Jabs at WebSphere 4.0 Microsoft Missing the Point, Says IBM Program Direct

WSJ-IN Exclusive: IBM Responds to Latest Microsoft Jabs at WebSphere 4.0 Microsoft Missing the Point, Says IBM Program Direct

Exclusive to Web Services Journal Industry Newsletter, here’s the latest fiery rebuttal in the red-hot IBM vs Microsoft debate that’s been raging on the Internet over the superiority of their respective platforms for creating Web services. Stefan Van Overtveldt, program director, WebSphere Technical Marketing, IBM, holds that Microsoft’s original white paper belittling WebSphere 4.0 was fatally flawed from the start due to its premise, which in his words, “is missing the point.” Before Van Overtveldt’s complete response, WSJ-IN reviews highlights of the verbal battle’s progress to date:

Microsoft launched the initial salvo in the Internet war, with a white paper that compared creation of Web services (using the PetStore.com scenario) using Visual Studio.NET versus IBM WebSphere v4.0. To support their claim that .NET has a significant advantage over WebSphere, Microsoft hired an independent consulting firm to develop a Web service using both its own platform and IBM’s. According to Microsoft, the results of this benchmarking exercise proved .NET the winner in developing Web services.

IBM responded with its WebSphere competitive review, (see ) www-3.ibm.com/software/info1/websphere/news/ibmnews
/compreview4.jsp?STACT=101CMW13&SCMP=campaign
calling Microsoft’s white paper “misleading,” and firmly stating that “there is no doubt that WebSphere is the superior platform for developing Web services. “ IBM pointed out that Microsoft’s study used the IBM Web services Toolkit, when the Web service should have more appropriately been built using the new WebSphere Studio Application Developer tool. IBM also said Microsoft overstated by nearly 6 hours the amount of time needed to create the service with IBM WebSphere. Furthermore, IBM said that .NET needed 106 lines of handcrafted code compared to 1 line in WebSphere Studio, and that total cost for constructing the Web service was lower with IBM.

Microsoft returned the volley with a “Response to IBM” (at www.gotdotnet.com/team/compare/ibmrespond.aspx ).

Microsoft’s reply minced no words: “IBM is attempting to mislead customers,” it states, in a point-by-point comparison of IBM’s claims and Microsoft’s positions on issues including cost of deploying the service, the number of lines of code needed to build and consume the Web service, the requirement of BizTalk server, the standards adhered to, and .NET’s ability to work in mixed environments. Microsoft did concede that IBM’s new WebSphere Studio Application Developer tool improves IBM’s support for building Web services, and Microsoft as a result issued a completely updated white paper comparing its product with that version (see http://www.gotdotnet.com/team/compare/
webservicecompare.aspx ). Microsoft has more background on its claims for the superiority of Visual .NET at http://msdn.microsoft.com/net/compare/default.asp.

For IBM's counter-reply, our News Desk spoke with Stefan Van Overtveldt, who had these comments:

Stefan Van Overtveldt: The premise of starting off with saying, “Let me show you how much more productive our development tools are by recreating the PetStore.com application in C# and with Visual Studio .NET,” is besides the point. PetStore.com is an application that been written to allow J2EE application vendors to test if all of the J2EE APIs are actually present and functioning well in their J2EE application server. That's the only objective that PetStore.com ever had. This being said, because it does go out and test all of those different APIs, this is an application that is not well written at all. It's not well written for performance. It's not well written for security. It is just a test. Taking this application and pretending that it's a real live application, and that the customers can draw conclusions with regards to productivity, performance, etc., is missing the point about what this application was intended to be.

When you look at trying to enable the application as a Web service not just from a development perspective but also from a deployment perspective (because, by the way, customers do want to deploy these Web services), we are still confident that we have tremendous productivity gains over any other application developer in the industry. If you look at what Web services are really all about, they're not just about just taking an application and making it available through an XML SOAP interface or to a WSDL wrapper, etc. You have to look into what is this going to do towards your entire infrastructure, what's the impact of opening up an application to the outside world, which is basically what you're doing. The impact is that you need to put in stronger security mechanisms. You need to be sure you can handle the workload. You need to be sure that you can quickly leverage these Web services as part of existing applications or expose existing applications as Web services. It's a much bigger picture than just taking an application and publishing it out as Web services.

You can debate different ways of doing this, but there's only one company right now that offers a complete Web services infrastructure, and that's IBM. We have Web services supported in our application servers. We have it in our development tools. We have a Web services infrastructure with regards to private UDDI gateways, for example, or private UDDI registries and UDDI gateway functionality. We have ways of managing Web services in a secure environment with a product like Tivoli Manager for Web services. We even have a number of technologies on the table that customers can use to take existing applications, not just stuff that was newly developed in C# or J2EE, but existing applications, applications that they've been using for years. Things like their SAP ERP systems, database applications, transactional applications running on IBM mainframe platforms, etc., and make those available very rapidly as Web services, again in the same managed and controlled environment. That is something no other company can offer.

WSJ-IN: The use of the PetStore setup did seem problematic.

SVO: We can argue about which company or which tool is more productive in this scenario for years to come. Truth is, it's beside the point. It's not a representative application. It does not make a lot of sense to go out and recreate an application in another language and see if it performs better. If I were to rewrite this application, and believe me I'm not a good programmer, it would probably perform better.

WSJ-IN: One issue aside from performance that Microsoft was slamming IBM on was cost. IBM's first response to Microsoft said a 4-server deployment using Microsoft .NET would cost about $399,996. Microsoft's response to that was IBM "is just wrong. IBM is attempting to mislead customers." Microsoft went on to say that WebSphere 4.0 costs over 15 times more that Microsoft .NET for this typical clustered scenario.

SVO: Well, if nobody is accessing the cluster, they're probably right. The WebSphere licensing is based on buying an application server, and we have a per processor licensing scheme, and you can put those applications, it's $12,000 list price per processor, in a cluster. So, if you have a 4-node cluster, for example, that would cost you again at list price, $48,000. Throw in 10 developer seats around and you're looking at a total solution of around $80,000. But that is it. Whether that solution is serving one concurrent user or it's serving 100,000 concurrent users, the price remains the same. We don't have the notion of access licensing. If you count how the developer license works, if you count on actually putting those applications in production, not just showing a demo of it, you have hundreds, thousands of other applications out on the Internet accessing that Web service, then the Microsoft solution is a lot more expensive. It just because it's a completely different way of counting licenses.

WSJ-IN: That's what is accounting for the cost differential: how they're approaching their cost basis?

SVO: This is the same thing that goes back to Pet Store. Pet Store is a lab type of application, something you would never put in production. And that has a number of consequences. If you set up this type of environment, in a lab environment, to calculate what it costs you, maybe Microsoft has a point. Maybe they don't. But if you put it into a production environment where you have to start accounting for a number of concurrent accesses to these servers, we are very sure that our solution is much more cost effective.

WSJ-IN: The last point Microsoft emphasized in their most recent response had been regarding code: Microsoft said that to create a PetStore Web service, IBM WebSphere Studio Application Developer required 82 handcrafted lines of code and Visual Studio .NET required 48, about half; and to create simple client to consume Pet Store Web service, IBM required 49 lines of handcrafted code, Visual Studio .NET24. Now is that still a function of Pet Store's being atypical of a real-life application code-writing situation?

SVO: Again, I did not personally run those tests. The one point I can make is that most of the ability we have in Web Studio Application Developer is to take an existing J2EE application and render it as Web services fully automatic. Now is there some hand coding involved and is it 48/24 lines of code, double in our scenario? I don't know. What I do see is that taking this application and making it available as Web services is, again, only part of what you need to do. Because you need to link this application to a security environment, which Pet Store does not do, by the way. You need to link this application to a management environment, which Pet Store does not do. With WebSphere Studio Application Developer we can do all of those things pretty much in an automated fashion. You may need to write some line of code here and there, but that is a much more realistic scenario to look at. I am absolutely convinced that if you look at what are the real overall requirements that you need actually to put a Web service in production our development tools are much more effective and much more productive than any other tool out there.

The larger question is, which platform are customers choosing to deploy Web services? Giga Information Group just issued a report saying that J2EE platforms are the big winners overall among early adopters of Web services technology and the most important to Web services strategy. When it comes to companies that are actually evaluating and deploying this, Giga says IBM WebSphere is the clear favorite over Microsoft. In the end, it is the market that is deciding, and according to Giga, customers favor WebSphere over .Net.

More Stories By SOA News Desk

SOA World Magazine News Desk trawls the world of distributed computing and SOA-related developments for the latest word on technologies, standards, products, and services and brings key information to you in a timely and convenient summary form.

Comments (1) View Comments

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.


Most Recent Comments
Glenn Bell 01/18/02 12:15:00 AM EST

It is really amazing how far and how low a marketing battle can go. Afterall it is just a marketing battle as no serious software vendor could believe simply re-writting an application in a differnet language gives you the ability to compare them in every possible way.

To add interest to the debate I spotted the following article.

http://otn.oracle.com/tech/java/oc4j/pdf/9ias_net_bench.pdf