Discussions

News: Gartner compares: How Long Can BEA Survive Against IBM?

  1. The JDJ discusses the up-coming Gartner report on the state of the application server market place. The article talks about the trends, and tries to guess at what will happen.

    Summary
    The annual Gartner Research Report analyzing the application server market share among leading vendors is expected to be out this month. Open-source JBoss has used its technical and business innovation in the J2EE app server market to take on the software industry giants, but the main race remains between BEA and a Big Blue gorilla 30 times its size - IBM's market cap as of April 2004 was roughly $150 billion vs BEA at around $5 billion. Who'll win...and who'll lose?
    NOTE: Another caution that needs to be taken into consideration while analyzing market share is that this information is extracted from the companies' financial reports and determined as their actual application-server sales. There may not be a perfect way to know market share shifts other than what they provide.

    These numbers always have to be taken with a grain of salt. There are many ways to book the numbers, etc etc.

    Read How Long Can BEA Survive Against IBM?

    Threaded Messages (40)

  2. Weblogic is great product and we all know app servers are not selected based on tech merits alone. Hope BEA survives against IBM.
  3. BEA will do fine[ Go to top ]

    Company size and capital value is not the deciding factor to determine a company's future. Otherwise, we probably should bet on Microsoft. :-)
  4. IMHO IBM has seen its share increase due to more and more big IBM shops starting to use application server technology. As far as I know BEA Weblogic is still the leader in new business opportunities (Unless someone knows different).

    We just migrated from BEA Weblogic 6.1 to 8.1 and it went pretty well. The new product is as you would expect from BEA easy to develop with and use in production. We did however drop WLI and Portal as the migration effort was going to be way too painful for these applications.
  5. wls6.1 + portal -> 8.1[ Go to top ]

    could you elaborate on your experience? we are in the process of evaluating 6.1+portal -> 8.1 path
  6. Most of the shops that I see running Websphere or OAS are doing so becasue they were given the product free based on past partnerships. Whenever I've seen a shop given the freedom to choose their J2EE container based on merits, they almost always choose Weblogic. The product is superior and that will always guarantee their survial, but you know that a superior product doesn't always translate into marketshare. Just look at your desktop OS!
  7. I don't think BEA is at the point of giving away licenses to get the business. While the app server technologies are similar, in that if you know J2EE you can develop in either, the marketed "products" are completely different. If you compare sales and marketing, services, support, training, and synergies with other products, I would think IBM has a higher cost of ownership because you need more services, support, and training with their products. They can make it back in aggressive sales (leveraging existing customer base), better marketing (lower cost license deals), and first class relationship with some third party products. BEA, on the other hand, is less of a threat to third party vendors (no DB2 shoved down your throat) so integration is great. For some reason WebLogic consistently works out of the box unlike the manadatory Global Services engagement to install and configure a cluster while SAs head off to admin training.

    BEA will survive as long as they listen to customers and are responsive to refactoring the feature set and architecture of the product to meet the IT direction of their biggest customers. If I were a Fortune 50, I would probably feel I could influence BEA to add a feature or cut a patch more so than I could IBM. That's control and a selling point. If you look at BEA's Thoreauvian direction to "simplify simplify simplify" Workshop, installs, admin, and documentation I thank thats a great focus and something IBM just can't seem to do. Check out the size of IBM's Redbook to upgrade app server version if you can find it.

    The New York Times had an article recently about offshoring not working out and coming home. While the offshore costs are a third, the productivity of onshore and in-house workers offset that difference. I think WebLogic's grade on productivity is a bit better than Webshphere at this point. I feel more productive with BEA tools, I have more confidence that they will work as marketed, and I know I can come up to speed on a service, feature, or configuration setting faster.
  8. Survive??[ Go to top ]

    I find it quite perverse to believe that a company has to vanish just because there is one or the other competitor. Without competition in the market place we would still be hacking away on cobol mvs boxes using punchcars (if at all). Even if the competitor is big, there is no reason, you should not excel by improving and reengineering your product.....

    And I quite like BEA WebLogic, its management console, the productivity and flexibility I get from it and most of the stuff BEA has based on WebLogic.

    Karl
  9. As long as they cost-effectively solve their customers' problems, they'll do fine.

    Ironically, Gartner is apparently still cost-effectively solving their own customers' problems.

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Clustered JCache for Grid Computing!
  10. Its the BUNDLE dude.[ Go to top ]

    Cost Effectively is a word then can be used by BEA and IBM . both maybe correct in certain ways in their perspective.

    The questions:
    1. Size of the pie.
    2. What markets/customer does BEA attack versus IBM.
    3. Mostimportantly , what is the Bundle
  11. IMHO IBM is not going to run BEA out of business. I don't even believe it's a priority for them. They are too focussed on doing a good job, too secure and confident to worry about driving other companies out of business.

    It reminds of the story in Pfister's book about when IBM published benchmarks on Parallel Sysplex. They used a fully shared workload (all nodes read and write the same data), and they said "the overhead is only 15%", or thereabout. If you know what the system is doing, you understand how good IBM is, but the press only heard "overhead". In this case IBM was too self-confident to do some decent marketing.

    These research firms are like political and sports commentators: they _have_ to start controversies or they are not happy.
  12. WebLogic vs. BEA[ Go to top ]

    WebLogic is a solid product, and has a fantastic customer base. They have done a good job of evolving their offerings as the commoditization to some degree of the application server has occurred. WebLogic is BEA's mothership. IBM, on the other hand, has WebSphere as just one of their many product and service offerings. With the partnerships that BEA has with HP and Sun, I would not be surprised to see them acquired somewhere down the road.
    Either way, WebLogic will continue to be around for many years to come I believe.
  13. Actually, I believe that BEA should look out for Oracle who has made moves to take over BEAs market share.

    BTW, anyone using Oracle's lastest AppServer? My company is evaluating the posibility of partnering with Oracle and dropping BEA.

    CarlosSA
  14. Stick with BEA if you can[ Go to top ]

    The list of issues I had with Oracle's application server are to long to mention. The most offending issue was that the Enterprise Manager would only function till I restarted the machine. It would not start anymore, and the only option I had was to reinstall it. Did I mention that reinstalling Oracle stuff is a major pain in the ... itself?

    I would happily choose Websphere over Oracle 9i AS R2 any day, without thinking twice. But I am not a big fan of Websphere either, to be honest.

    I just love working with Weblogic, however. :)
  15. BEA is best, but...[ Go to top ]

    I have delivered applications on weblogic 8.1, Oracle 9.0.3 and Websphere 5.1 and Weblogic was by far the best experience for me. Everything just works the way it should and the documentation is pretty good.

    Oracle was a pain to live with, the documentation was all over the place and I got the feeling that the product had been glued together from a bunch of separate technologies, by a bunch of separate teams, and no-one talked to each other.

    Websphere + WSAD was very productive, but the learning curve was quite steep - Getting CMP working was a MAJOR pain.

    I really hope BEA can hold out, but to be honest I'm backing IBM. - At the end of the day the choice of app servers is a management decision, and in the long term I think it's going to be far easier for your average manager to justify the decision to go with IBM.
  16. Oracle vs BEA[ Go to top ]

    Don't do it! I have used WebLogic a lot in the past, and whilst I've had problems with it it was generally good to work with. In my current job I'm using Oracle AS (9.0.3 and 9.0.4), and it's driving me crazy - development with OC4J isn't too bad, but the management tools (Enterprise Manager Website, DCM etc.) are very unreliable, especially if you're using clusters. Also, the recent release dates for Oracle AS have slipped by a ridiculous amount; 9.0.4 was due out "in 2 months" for about a year! Have a look at the forums at http://forums.oracle.com/ for details.
  17. These research firms are like political and sports commentators: they _have_ to start controversies or they are not happy.
    I'd go a step further to say they're similar to the media. Controversies and attractive titles (or "creating interest") make them get paid.

    If one wants to get real opinion about products, one has to tune in to the forums, user groups, etc. "We" are the users, "we" drive the industry...
  18. only 5 billion[ Go to top ]

    Gee, BEA only has 5 billion to survive on, I wonder if they will last ... With the talent they have they could re-invent themselves several times over, but they've done such a good job they don't really need to. Based on the changes in Websphere, I think it is more accurate to say the BEA has been lighting the fires under IBM.
  19. only 5 billion[ Go to top ]

    Gee, BEA only has 5 billion to survive on,...
    BEA's market cap is not BEA's cash stash. Market cap is the worth of the company as decided by the shareholders and share price. It currently stands at $3.4B, down sharply from around $5B before BEA's earnings call, where they announced they had missed their software license revenue targets. While BEA is basically flat, and the execs are blaming "tough market conditions" and "longer sales cycles". At the same time IBM's John Swainson is saying WebSphere had its 22nd consecutive quarter of growth (in an interview in the same issue of JDJ). WebSphere grew 24% in 1q2004.

    I am approaching this not as a technologist but as an investor.

    The next interesting question is, how long will it take for Oracle to buy BEA. Oracle just dropped their offer for peoplesoft by $2B. BEA has $1.5B in cash. So with the $2b they saved on PSFT, and the $1.5B BEA has, Oracle can purchase BEA for "nothing". As soon as Larry comes in off the yacht, look for the news headlines.

    and the obligatory disclaimer: I work for Microsoft. !
    and these opinions are my own.
  20. An interesting argument I heard in an internal architectural debate was that since the app server is standards based and conceptually mature it is now a commodity. This group ended up going with BEA because of the suite - they needed the integration and portal and BEA provided all three (IBM didn't really have the suite together at that time).
    I suspect these "commodity" arguments may become more common. Empirically, I do not agree with them - I think BEA is easy to use, has superior EJB support while IBM with MQSeries is probably still ahead in messaging.
  21. An interesting argument I heard in an internal architectural debate was that since the app server is standards based and conceptually mature it is now a commodity.
    It might might be an interesting argument but it's wrong :-)

    There are absolutely no signs whatsoever that appservers are becoming a commodity. Some of these signs would be:

    - Prices driven down regularly
    - Hardly any distinction in features between all the servers
    - Dozens of players sharing the market, each having a fragment of it

    Toasters and VCR's are a commodity. Appservers... not just yet.

    --
    Cedric
  22. Toasters and VCR's are a commodity. Appservers... not just yet.
    I don't even know about that ... any market in which a brand decision overrides a price decision can't be a commodity market, can it?

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Clustered JCache for Grid Computing!
  23. J2EE space: Commodiized?[ Go to top ]

    Toasters and VCR's are a commodity. Appservers... not just yet.
    I don't even know about that ... any market in which a brand decision overrides a price decision can't be a commodity market, can it?
    Apart from (or prior to, rather) the brand neutrality, the product neutrality (no differentiators) is an imp aspect of commoditisation.

    In the J2EE space the product functional neutrality is more or less there. Among certified servers, a common level of functionality (as seen by the application and app developers) is assured. Prob one of the first app platforms where it is actually possible to move from one vendor to the other. So, remaining differentiators are primarily on management (an USP in Pramati's case) and perforamance and scalability.

    Surely J2EE is a space rapidly heading towards commoditization. (Even if one of the final attributes of commoditization may never be reached, which is- 'barriers to entry for new players should be very low'!).

    Cheers,
    Ramesh
  24. I agree as well but I think many of the Gartner influenced IT architecture departments may be thinking that way.
  25. Commodity, etc[ Go to top ]

    There are absolutely no signs whatsoever that appservers are becoming a commodity. Some of these signs would be:- Prices driven down regularly- Hardly any distinction in features between all the servers- Dozens of players sharing the market, each having a fragment of it. Toasters and VCR's are a commodity. Appservers... not just yet.
    Guess the slightly overused analogy with databases still applies. The market is fairly stable, there's a few main players (MS, ORCL, IBM), a bunch of also-runs with insignificant market share and a couple of free offerings. The prices for the leading products aren't low by any means. Can we call it commoditisation? Perhaps... In any case, it's pretty good for the market leaders and not too bad for their customers either.
  26. IBM in your shop = Consultants gone wild!!!

    Yes, you too can own the video!!! Only $19.95!

    VJ
  27. Every time IBM releases a new version of WebSphere the complexity of the product increases ten fold. They will tell you that they are providing more functionality but in fact, they are taking a very simple product and making it difficult so that they can come back and sell you consulting services. Take for example WebSphere 5.0 and Network Deployment. We put the enterprise edition on our servers only to find that when creating multiple server instances, the admin console will only administer apps on the default instance. In order to manage multiple instances you need to install an additional product (Network Deployment) which of course isn't free. Even more frustrating is that the info center documentation clearly states that the admin console can administer multiple apps on multiple server instances. In actuality, without ND you would have to install the admin console on each server instance (using a different virtual host, of course).

    BEA put out a superior product in my opinion and should not be threatened by IBM.
  28. WebSphere packaging[ Go to top ]

    Take for example WebSphere 5.0 and Network Deployment. We put the enterprise edition on our servers only to find that when creating multiple server instances, the admin console will only administer apps on the default instance. In order to manage multiple instances you need to install an additional product (Network Deployment) which of course isn't free.
    I'd like to add some clarification to these comments about WebSphere packaging. If you bought WebSphere Enterprise, then you have Network Deployment. It is a part of Enterprise. You just install it and it federates multiple servers into a cell so they can be managed as a single group. For configurations where only a single server is needed, Network Deployment is not needed, so it does not need to be purchased or installed. For more complex environments with multiple servers, multiple applications and specific workload management requirements, Network Deployment is the component that will provide the system management functions needed to deploy applications to multiple servers, provide workload management and administration.
    This all about allowing clients to purchase and install only what they need, and it is all very publicly documented.
    Yes, I work for IBM (but the opinions are my own etc )
  29. WebSphere packaging[ Go to top ]

    Take for example WebSphere 5.0 and Network Deployment. We put the enterprise edition on our servers only to find that when creating multiple server instances, the admin console will only administer apps on the default instance. In order to manage multiple instances you need to install an additional product (Network Deployment) which of course isn't free.
    I'd like to add some clarification to these comments about WebSphere packaging. If you bought WebSphere Enterprise, then you have Network Deployment. It is a part of Enterprise. You just install it and it federates multiple servers into a cell so they can be managed as a single group. For configurations where only a single server is needed, Network Deployment is not needed, so it does not need to be purchased or installed. For more complex environments with multiple servers, multiple applications and specific workload management requirements, Network Deployment is the component that will provide the system management functions needed to deploy applications to multiple servers, provide workload management and administration.This all about allowing clients to purchase and install only what they need, and it is all very publicly documented.Yes, I work for IBM (but the opinions are my own etc )
    To be sure, ND does some things that are nice such as managing multiple nodes, etc. However, WAS 4.0 (as well as other popular J2EE servers) had the ability to manage multiple server instances on a node from the admin console. Have they left out the middle sized companies? I'm am either forced to buy enterprise with ND at an additional cost or I can only run with one server instance per node. Most of the people in my organization wish we had never upgraded and are now considering other app servers.

    But to be fair, we have managed to iron out all of the issues in our own way using wsadmin and automated deployment and now have few problems with our production environment. Still it took many frusting hours on the phone with incompetant tech support technicians to get this done.
  30. WebSphere packaging[ Go to top ]

    Take for example WebSphere 5.0 and Network Deployment. We put the enterprise edition on our servers only to find that when creating multiple server instances, the admin console will only administer apps on the default instance. In order to manage multiple instances you need to install an additional product (Network Deployment) which of course isn't free.
    I'd like to add some clarification to these comments about WebSphere packaging. If you bought WebSphere Enterprise, then you have Network Deployment. It is a part of Enterprise. You just install it and it federates multiple servers into a cell so they can be managed as a single group. For configurations where only a single server is needed, Network Deployment is not needed, so it does not need to be purchased or installed. For more complex environments with multiple servers, multiple applications and specific workload management requirements, Network Deployment is the component that will provide the system management functions needed to deploy applications to multiple servers, provide workload management and administration.This all about allowing clients to purchase and install only what they need, and it is all very publicly documented.Yes, I work for IBM (but the opinions are my own etc )
    To be sure, ND does some things that are nice such as managing multiple nodes, etc. However, WAS 4.0 (as well as other popular J2EE servers) had the ability to manage multiple server instances on a node from the admin console. Have they left out the middle sized companies? I'm am either forced to buy enterprise with ND at an additional cost or I can only run with one server instance per node. Most of the people in my organization wish we had never upgraded and are now considering other app servers.But to be fair, we have managed to iron out all of the issues in our own way using wsadmin and automated deployment and now have few problems with our production environment. Still it took many frusting hours on the phone with incompetant tech support technicians to get this done.
    I see now that my clarification was not very clear. My apologies.
    I don't know if anyone is still on this thread, but basically, the major versions of WebSphere 5.0 are
    1. Base (full J2EE 1.3 implementation but has no workload management)
    2. Network Deployment (ND)- Adds workload management, clustering etc. This product is comparable to WAS 4.0 Advanced which is the one you must have been using previously. This is the version most corporate deployments use.
    3. Enterprise - This package is a superset of ND, adding a set of programming model enhancements (advanced features that are beyond the scope of J2EE such as Asynch beans, business rules, process choreography and many others), and a few other features. It can be installed on top of either Base or ND. The Enterprise package includes ND however, so if you have purchased Enterprise, you do not need to make an additional purchase.

    So ND is a superset of Base, and Enterprise is a superset of ND. The point I was trying to make is, if you purchased WebSphere Enterprise, you do not need to ALSO purchase ND to get workload management, multiple servers etc.

    It sounds like you might have thought you needed to buy Enterprise to get work load management. That is not true. Hopefully this post better explains the packaging.

    The middle-sized are not left out. They would go from WAS 4.0 Advanced to WAS 5.0 ND and get the same (actually improved) support for high volume configurations with multiple servers and workload management.
  31. What's really in the report[ Go to top ]

    The article claimed it was written before the Gartner report was released and the gartner report should be out by now. Is it true? Anyway knows what's really in the report?
  32. What's really in the report?[ Go to top ]

    The article claimed it was written before the Gartner report was released and the gartner report should be out by now. Is it true? Anyway knows what's really in the report?
    I heard it was 41% IBM to 28% BEA for 2003, in the commercial J2EE app server space. This changed from 37/29 the year prior.

    No mention of JBoss, JoNas, Geronimo, etc.
  33. It's been interesting to watch this David and Goliath battle between BEA and IBM. I have to admit my bias before I say anything -- I love WebLogic and I have been using WebLogic since early 1999 and wouldn't want to use anything else. We currently have most of our apps in production running WebLogic 8.1 sp2 and it just rocks. WebLogic is a rock-solid appserver with great performance, scalability and the combination of WebLogic and JRockit makes it a very compelling solution.

    BEA is also on the leading edge when it comes to supporting the latest specs and jsrs. I hope BEA hangs in there with IBM in the numbers game as a lot of CIO/CTO's make decisions based on Gartner numbers and not actual technical evaluation.


    --Vinny
  34. How does .Net factor into all this[ Go to top ]

    Could the IBM and BEA market be facing competition from application server space outsiders like M$ via their .NET?

    Here I am talking about J2EE application servers being replaced by machines running Win2003 or win2k, IIS and .NET or is this not a practical thing?
  35. noise about nothing[ Go to top ]

    You know GM now has the largest car share in the world. I wonder how long Ford will survive...

    It's this sort of pointless logic that generates dollars for the analysts. You know in any industry their are always 3 or 4 main competitors. Look at the DB market. IBM has the largest share now. But does that mean Oracle will dissapear.

    Unfortunatley not :)
  36. WLS x WAS[ Go to top ]

    I wonder why would anyone move from BEA to any other app server: there is NO PRODUCT that compares with it, specially if you are pushing J2EE features to their limit.

    Most WAS users I have seen aren't using anything beyond servlets/JSPs (so WAS behaves just like a Tomcat with steroids for them). Trying to use CMPs in WAS (and in Oracle, as I have been told) is a plain waste of time. I have many stories to tell about how frustating this is.
  37. Weblogic and Borland Enterprise Server are the only real commercial viable application servers period.

    WebSphere like Oracle9iAS are for Pointy-haired driven IT shops. Its technically impossible to validate WebSphere, its purely a political choice.
     
    I wonder how many J2EE shops are using local EntityBeans as a foundation for their domain layer?
  38. I wonder how many J2EE shops are using local EntityBeans as a foundation for their domain layer?
    <raises-hand>

    Ohhh... did you mean non-BEA shops?

    <lowers-hand>

    cheers,
    Markus
  39. BEA as a company has been around for a long time (before J2EE appservers), and has a number of other enterprise products such as Tuxedo. Even with the WebLogic platform, the appserver is just the 'base' product, with other very useful products such as WebLogic Integration and WebLogic Portal. WebLogic Workshop is one of the leading solutions for easing J2EE development. There products seems to be well integrated as well. So my conclusion is that BEA will have a decent chance of fending off bids unless someone puts in a ridiculous one.

    Well, of course that is assuming you have the money and are willing to hedge your bets on their platform.

    ps: I don't work for BEA, have used their stuff in some external projects.
  40. The answer of how long they can survive is "as long as IT shops do objective feature-by-feature comparisons". I've done 3 side-by-side comparisons now and WebSphere has never even come close. However, as someone mentioned, there are 2 pretty good examples out there of how a powerful, well-funded marketing machine can acquire market dominance. So, the game's not over. It's just a matter of which giant decides to buy BEA. IMHO that's the most probable outcome - assimilation. BEA's sitting on some advanced technology right now that others just haven't come up with yet. They have managed to do a great job of creating an endless series of competitive technical advantages that have left their competitors always in a position of trying to catch up.
  41. WebLogic will survive[ Go to top ]

    Anyway, WebLogic is a great product. Even BEA can not survive against IBM, WebLogic will survive against WebSphere. There are so many job oppotinities of WebLogic, it means there are so many WebLogic running.

    Most customers using WebLogic are satisfied with it. And they will purchase new WebLogic licenses if needed. Not WebSphere.

    WebSphere is more used with financial services, manufactures. While WebLogic is more used in tele-communication services, middle-size companies.

    I believe WebLogic will survive only because it's really a great product. Even BEA dies, WebLogic will be survive with some other brand.