The rise and fall of App Server Commoditization

Discussions

News: The rise and fall of App Server Commoditization

  1. Felipe Leme has a blog entry entitled "The rise and fall of the application server as product (A.K.A. app server commoditization)," which questions whether commodization has actually happened or not - based on the number of adopters of each revision of the specification so far: from 1.3 to 1.4, licensees went from 29 to 13. Java EE has only two licensees so far (Sun and TMaxSoft), but considering the age of the specification, it's a little early to assume that it will only have two licensees total.
    I know it's too early to analyze the JavaEE 5 arena (I'm sure JBoss, IBM and WebLogic will come with their implementation as well), but the J2EE 1.4 numbers show the trend is true: the JSR-151 (Editor's note: the J2EE 1.4 JSR) specifiation has been out there for almost 3 years and the total number of servers available is less than 50% of its predecessor, J2EE 1.3 - not to mention that many servers took years to be ready.
    It's a worthwhile consideration, though. Wouldn't commodization imply more implementations and licensees rather than a dropping number? As Mr. Leme points out, the complexity of the specification may be part of the issue, but - again - commodization would imply building application servers from best-of-breed projects based on specific requirements, so the number would be expected to be higher. What do you think the conclusion should be, and why?
  2. ... ?????[ Go to top ]

    Sorry guys, but is this supposed to be a Blog entry ?? I mean, such a title with such a "non-content"....??
  3. The number of certified J2EE servers, or Java EE servers, is not an indicator for popularity of them. In some degree, that is an indicator of mature of J2EE sector. For every sector, at one time, there are many players, some time later, only a few remain. Of course, J2EE sector is not strong. Many users still use J2EE 1.3 (Weblogic 8.1, Websphere 5.x). Companies do as others do. In year 2000, if you buy new software, I do, if you hire new people, I will. But, now, if you do not buy new software, why should I? Many companies do have issues for their application servers or applications. They prefer manual work-around to buying/updating new software. If manual work needs more people, they hire new people. It is easy to ask budget for extra worker than to ask budget to buy new software. Wei Jiang Perfecting Java EE!
  4. I'm sorry, but doesn't that imply that commoditization is taking place? Commoditization IMHO, not only decreases the value of creating a new product in that space, but also decreases the value of existing products in that space. Thats why we have only Tomcat and Jetty in the webserver space and Eclipse, Netbeans and IntelliJ in the IDE space after commoditization.
  5. It's certainly trying to get commoditized. The struggle is that while appservers are getting their, the applications aren't quite there yet. While you can write an application that ports easily from one appserver to another, most folks I think do not. That limits the commodity status of the appserver itself. For example, Solaris now bundles SJAS 8 with the OS. Apple is (I believe) bundling JBoss with their OS. And, of course, both are readily available for most anything. But it's not clear yet that applications are truly app server agnostic. Web servers are in a similar boat. Yes, I can write a portable web application if I rely solely on CGI. If I rely on an Apache Modules, then it won't run on IIS or perhaps even Apache 2. Yet, everyone expects to have a web server. So, with JBoss and Glassfish/SJAS, there is little stopping J2EE from becoming commditized. Now we just need apps to be better tested across them.
  6. I've been watching this argument for years. I joined BEA in 1999, when the app server space took off with the convergence of Web apps, J2EE standards and the dot com boom. After the bubble burst BEA went into a period of flat to slightly negative license revenue, which some argued was fueled by free and low cost alternatives eating into the space a.k.a. “the commoditization of the app server”. I think this theory has been greatly over blown. Here are a few data points about BEA’s core business to consider; It is interesting to look at the number of transactions we’ve been doing each quarter. In Q1 of this year, BEA did 2,451 license transactions, up about 10% from last year. Our big deal count (over $1M) was about the same as Q1 last year, so the theory that JBoss, Oracle and others are cannibalizing BEA's business, particularly at the low/mid range, simply doesn’t hold up in the numbers. After the bubble burst, BEA did have a period of flat to slightly negative license revenues. This fueled the commoditization argument. But last year, as IT spending stabilized, BEA returned to y/y license revenue growth in Q2, accelerated it in Q3 and finished the year with 18% y/y license growth. We carried that double digit license growth into Q1 of this year, so again, this theory that app servers are interchangeable so you should use the lowest cost one simply isn’t how the market is behaving. I’m throwing these data points out there to get some of the religion out of this argument and to bring it back to real business facts. Flame on :-) Eric BEA Systems
  7. No flames or religion here. My preferred J2EE server is Weblogic. But the arguments you give are the same ones that JBuilder could've given a few years ago against the commoditization of the IDE space. And where are they today?
    I’m throwing these data points out there to get some of the religion out of this argument and to bring it back to real business facts.

    Flame on :-)

    Eric
    BEA Systems
  8. Nilesh, I guess I would say that each product space has different dynamics. Eclipse did a number on the IDE market but wasn’t the only reason Borland faltered. The fact that all of the platform players provide free tools makes that space very difficult to compete in. On the other end of the spectrum, look at the RDBMS space. There have been free open source products in that space for years, yet Oracle still runs a multi-billion dollar business around it. I agree with Michael's position above. The barriers to entry are high in this space. For some use cases free open source products are fine. For others, it isn't. Eric
  9. JBuilder is in clearly in a weak position today. Had they seen the train , they would have jumped onboard the Eclipse train, ported their gee-wiz features on top of it, and they could have ridden the wave. Instead they have Peloton, an eclipse-based preannouncement that's TBD later in 2006 or 2007. Where there's value, there are customers. IntelliJ is still kicking around with a loyal base. M7 NitroX, now BEA Workshop Studio is doing ok. Relational databases were suppose to be commoditized in the 90s and yet - not so. Water should be a commodity and yet, many times we pay $1+ for filtered tap water that's free. Salt - now that's commoditized. Unless you want sea salt, which costs more. Or maybe organic salt. Isn't marketing and positioning cool? ;-P
  10. JBuilder is in clearly in a weak position today. Had they seen the train , they would have jumped onboard the Eclipse train, ported their gee-wiz features on top of it, and they could have ridden the wave. Instead they have Peloton, an eclipse-based preannouncement that's TBD later in 2006 or 2007.
    The rise and fall of JBuilder... We dropped JBuilder in our company when Eclipse started to just have more going on with it and its plugins. It seemed like Borland couldn't really come up with any new features that were actually useful. Instead they shortened their release cycle and we ended up with upgrading to a new JBuilder versions, only to find out that the previous problems were still there. But to be fair, Eclipse and all its plugins (including the ones that can be considered as complete as a full blown IDE) have problems that were better handled by JBuilder. One of these is for example the group working aspects and working with project groups in general. The workspace concept just doesn't gel well with me... BTW, Peloton in addition to the French reference means fearless in Finnish. So you should be "peloton" to use Peloton... ;) -I
  11. I've been watching this argument for years. I joined BEA in 1999, when the app server space took off with the convergence of Web apps, J2EE standards and the dot com boom.

    After the bubble burst BEA went into a period of flat to slightly negative license revenue, which some argued was fueled by free and low cost alternatives eating into the space a.k.a. “the commoditization of the app server”.
    Has BEA improved it's licensing model? When we evaluated it, it was the number one choice for quality but charging the exact same (high) price for each instance in a cluster seemed pretty asinine. The Websphere model seemed much more appealing.
  12. Mixed Metaphors[ Go to top ]

    Commodization is a well known concept in economics. From a pure economics perspective, the author's data points do not indicate commodization. If there was commodization, then many vendors would offer app servers that were as good (or "nearly" as good) as those offered by entrenched vendors such as BEA and IBM, but they would be offering them at lower prices to attract business. Less app server offerings would indicate that other vendors cannot offer the same functionality easily. But there are a lot more factors going on here. This is not the first time I've seen open source software and commodization mentioned at the same time. They seem like similar concepts. Open source software makes a product free, and commodization makes products price drop towards cost. They really are different things though. In fact in some ways open source software protects the app server market from commodization. If I'm a software vendor and I can make an app server that is as good BEA and IBM's for a lower price, I would do it, right? Not if there is open source alternatives. With open source alternatives, I will have a hard time winning any pure-price comparisons since my app server is not going to be free. I need to be able to beat IBM and BEA on features, since I cannot beat OSS on price.
  13. Whatever you call it[ Go to top ]

    I think the answer depend son how you define commoditization. Most people would agree on the facts: - an app server is something you expect to see as part of a complete system delivery - there are fewer app servers on the market - it is becoming increasingly difficult to get customers to pay for the commercial app servers due to the fact that the free ones are actually very good Regarding the fewer brands of app servers, then that is very natural. It has happended in a lot of markets. In the industrial history: cars. In recent history: CPU's. I believe that many people consider the x86'ifying of the IT world for commoditization even though the number if CPU's out there are decreasing.
  14. Re: Whatever you call it[ Go to top ]

    The original author is NOT taking about commoditization he's talking about market consolidation - or he has no idea what commoditization is. It doesn't seem unreasonable that the number of vendors has decreased (it was pretty high to start with) - this shouldn't be seen as a reflection of the market's strength. That aside I think segments of the market commoditized a couple of years ago - and as a whole - Java EE is now in the "late majority" wave of growth where convenience and price are the dominant criteria for choice. Any technology that is based on Open Standards will create conditions where commoditization is likely. This is a natural evolution of a technology - nothing to see here - move on. I firmly believe that the major market growth for Java EE (and Java) is in front of us; not behind. Rich Sharples Sun Microsystems http://blogs.sun.com/theaquarium
  15. Re: Whatever you call it[ Go to top ]

    Gartner just published new app server market share data. For those of you who track this stuff, the report can be found here; http://www.gartner.com/DisplayDocument?ref=g_search&id=493345 What we are seeing is market consolidation, not commoditization. Eric
  16. Re: Whatever you call it[ Go to top ]

    Consolidation is a sign of commoditization. When manufacturers can't differentiate or position their product in a unique way it becomes a commodity, reducing its price and leading to consolidation because small players can not make a profit under such circumstances.
  17. Java EE market slowing[ Go to top ]

    I can't help but wonder what would this thread look like if the author had used those licensing numbers to conclude Java EE market is shrinking or app server market is slowing down or something like that. It would have been a lot of fun that way:) C http://ChintanRajyaguru.com
  18. I do agree when it is said the app server is commoditized. There was a time when there were only a few. Weblogic, JBoss, JBuilder. These have maintained their stride consistently. They stood for great performance and reliablility. But now it seems every other Jack and Harry is building a app server out there. A case in question is of a recent article: http://www.theserverside.com/news/thread.tss?thread_id=40968 of a product called Jeus by a company called Tmaxsoft, which it seems has got Java EE certified. Some cause for celebration. A app server is known by its ability to service and store requests dynamically. Compatibility has little to do with it. But nowadays if u have to sell your product and beat the competition just flash lovely add ons, if u dont want to invest funds for research. I think guys it time to evaluate the real good ones and keep aside the gimickery of new ones which promise more but deliver less. A true waste of time and energy.