Discussions

News: AMR: App-Server Market To Boil Down To Four Survivors

  1. AMR Research said Thursday that application-server vendors had for the most part reached feature parity, and predicted the $3 billion market would come down to four significant players: BEA Systems, IBM, Microsoft, and Oracle. The report also mentions that existing infrastructure and IT skills will influence platform choice — not application server features and functionality.

    Read AMR: App-Server Market To Boil Down To Four Survivors

    Read AMR's Press Release.

    Report Highlights

     - The role of the application server is changing and becoming a key part of the overall infrastructure. Vendors are creating application infrastructure portfolios that include portals, integration, business intelligence, caching, personalization, and development tools.

     - Interviews reveal that application servers create real technology value in the areas of employee self-service, Web-enabled applications, better visibility into the supply chain, information aggregation, and B2B collaboration.

     - Often ignored in application server analysis, research shows that Microsoft is moving towards becoming the third largest application server provider, and will continue to increase momentum due to a solid .NET strategy.

     - Selection between Microsoft and J2EE vendors depends on existing infrastructure and IT skills—not on application server features and functionality.

     - Some vendors are using a new breed of easy to use development tools to differentiate their application servers allowing application deployment to their company's application server.

     - Integration is still a major pain point, with certain vendors building the technology to address this issue.
    For users, the only way to verify vendor's multiple claims is to engage vendors in a prototype project, which allows testing for development, deployment, and performance of a real application component.

    Threaded Messages (16)

  2. Sorry, but I can't second this, because of the following:
    *) While the "researchers" (Gartner, AMR, ...) predicted that only a few vendors will survive there are still plenty of them (see http://www.flashline.com/components/appservermatrix.jsp), I even have the impression there are now more than ever...
    *) They already mentioned "wildcards" like JBoss (Tomcat IMO is no AppServer), maybe Jonas too
    *) Borland seems to generate enough revenue from their AppServer, I don't think they just "hope for the future"... From my impressions, they currently _gain_ market share; BES was also chosen as the second (behind BEA) best appserver in some readers choice awards (JDJ?); They also have a strong user-base with JBuilder and VisiBroker
    *) HP has just "started" their HP-AS and it looks great, there are also great features to come with the next releases... they invested lots of money and "rejected" BEA some days ago, so I don't think they will immediately "stop" HPAS (and its a great server too)
    *) Not sure about Iona and Silverstream, but they have a small chance I'd say...
    *) Hitachi Cosminexus seems to be the #1 player in the Japanese market... and I think this market is big enough to survive easily ;-)
    *) iPlanet: Okay, I'll never understand Suns software strategy and while I think they just "cannot" give it up they always surprise me with strange moves and failed software strategies
    *) Macromedia seems to be strong with ColdFusion and Flash and has a big user-base... and their server does J2EE too, though I don't know where they'll be going...
    *) Pramati has just started, so why should they suddenly stop?
    *) Sybase also has quite a user base and made quite some investments in their AppServer, they also need it for PowerBuilder-Apps I think. And it really has some unique abilities
    *) Trifork is also a new player, don't think they'll immediately give up

    So while I agree there will be a consolidation, some vendors will cease to exist I don't think it boils down to only four... freedom of choice will remain I guess, though I agree the major players will be BEA and IBM... not so sure about Oracle, why should anyone use their server??? (besides "pure Oracle shops"... I know few of them, as just a database is obviously not enough to run a business)

    Any thoughts?

    Messi
  3. There will probably be a many of J2EE application server vendors in the future, but the majority of the revenue will be taken by less than a handful of vendors. As a rule of thumb the top 3-4 vendors in any segment usually account for more than 80 percent of the revenue in that segment. Just look at database and ERP systems where 2-3 vendors completely dominate.

    I'm sure that companies like Borland and HP will continue to offer application servers and there will also be numerous open-source alternatives, but their revenues will be dwarfed by the revenues of BEA and IBM.


  4. i would not rule out atg - quite an install base
    apart from the sun - thing iplanet i think all the rest will come packaged with services . sun's java stategy is always obscure so i dunno there but my bet is on bea ibm
    atg microsoft
  5. ATG is a piece of crap
  6. ATG is already banking on not being a pure app server vendor -- that's why they've started moving to offering their products (like personalization services) on other app servers. Since those "value add" products seem like the main reason people buy ATG, moving out of the app server market and supporting lots of other vendor app servers is probably a good move for them.

    Mark
  7. ATG has concentrated on solving real world problems for our
    customers. In the beginning this amounted to a better alternative to CGI, resulting in the Dynamo Application Server (DAS). Over the years the problems our customers have needed to solve have become more and more complex, giving us the opportunity to provide applications and
    services centered around personalization, relationship management and commerce.

    Throughout this history one of ATG's continuing goals for technology design has been centered around Enterprise Fit. Organizations standardize on hardware platforms, they standardize on operating systems, databases and now application servers. ATG has gone to great lengths to ensure that our products support the enterprise components
    that dominate our customers' IT infrastructures. We run on the leading hardware and operating systems, we support the leading databases, we integrate with a wide array of other enterprise applications. This has logically extended to support other J2EE application servers. Our existing customer base and market demand were the primary influencers in our decision to support BEA Weblogic and IBM WebSphere first. ATG will support other J2EE servers as customer and market demand dictate.

    It is our belief that companies want to standardize on an enterprise platform, which includes a variety of hardware platforms, operating systems, databases and application servers. From our perspective we want to be the standard online CRM application platform. Though we are committed to DAS, it is in the best interest of our current and future
    customers to become more agnostic to such low level plumbing.

    It should be noted that ATG is fully committed to the continued development of DAS, with full J2EE 1.3 support later this year. DAS is optimized for high traffic personalization and commerce sites. We expect that DAS will continue to be one of the optimal application servers for
    our typical customers' applications.

    Bob Mason
    bob at atg dot com
    ATG Senior Software Architect
  8. Thomas

    I agree that 80% revenue over a four way split is a fair way to look at things but there are other factors involved, such as size of installed user base, the 'I-pay-so-much-annual-support-but-get-no-new-features' issues, the actual vendor infighting itself. If you look at these being the 4 big players because they publish the highest quarterly profits to Wall Street, then ok.

    However, JBoss and Jonas, at the J2EE end, Mono and DotGNU at the .NET end and other techs like Apache, Jini and IBM's look at adaptive networks, will greatly skew these highly analytical 'findings' somewhat

    Just because a company has market share does not dictate that it has mindshare.

    Even though you are probably right in that eventually it will all come down to money

    Calum
  9. I agree completely,

    noone is saying that all the little guys will go away - although many of them will - but the market will become more and more of an oligopoly, as you reightly pointed out is already the case with the OS and DB market.

    It gets harder to break into the app serv. market and establish yourself without a huge investment (as shown by Oracle App Server) - Watch Oracle - they are aggressive.

    But another interesting facet is that SAP & Siebel - the two biggest app vendors in the world - have their own application servers as part of the infrastructure.....
  10. I didn't know MS had an Application server. IIS and COM+ (even with .Net tossed in) really don't compare to any of the Java based offerings.

  11. They essentially do, if you factor in all the different pieces they offer. They have stuff like ASP/ASP.NET, transaction services, message queue, components (COM/COM+/.Net), object pooling, connection pooling, etc. So, they have lots of the features you traditionally get in an app server, it just isn't really packaged up as such.

    Mark
  12. This report is FUD pure and simple - I wonder which of the four big companies it was sponsored by?

    How they can not consider THE MARKET LEADER (ie JBOSS) is beyond compression.

    Ignore this report - the authors obviously live on a different planet.
  13. Having used both MS and Java products (MS more than Java) they just don't seem to be that close. MS's products seem to exist to just support a GUI.

    With the addition of .Net it comes closer though. I hate MTS(whatever it is called in COM+ now). COM+ has made it easier but I still have to know I'm doing COM+. In java I can create my persistance and not worry if it is serverside or not. Unless I'm doing EJBs - which are overkill for most of my apps. But at least I have a choice.
  14. IMHO, the survival of the App Server companies will not relate to what we currently see as the app server, but to something larger and associated to it. App Servers are becoming more and more commoditized - where's the margin in a $500 product that is not consumer scale? Or in a $0 product (such as JBoss, which probably has the biggest deployment share).

    Their survival depends on other stuff. MicroSoft has the OS; IBM has the hardware, the OS, the DB and services; Oracle has the DB, the apps and services; Sun has the hardware, the OS and services. HP has the hardware, the OS and services. What does poor old BEA have??

    I'd say unless BEA branches out, it'll come down to IBM, Oracle, Microsoft and Sun/HPaQ
  15. BEA has great engineers who produce great software, that works on all those hardware platforms, and on all thoses OS's.

    I say they're sitting pretty.

    BEA want's to obviate services with simple to use, expressive and powerful products.

    IBM, to a great degree, relies on services, and so do their customers.
  16. what is in J2EE what you will not find under MS?
  17. the freedom to choose the best of breed available...not the monopoly