Short of openly admitting various problems (difficult to install, lack of stability, performance, and scalibility) with its own J2EE products offerings (OJSP, OSE, and EJE), Oracle quitely switched focus to OC4J, which is Orion Server with an Oracle label. Oracle also recommended customers not to use any of the J2EE containers it offered up to now for J2EE deployment. For details, see:
Read all about the new Oracle here
click on the FAQ section, search for orion
Will this move help Oracle to gain some market share in the app server space?
Hmm.... so Oracle is becoming the micro$oft of the y2k even more by starting to steal stuff from people and calling it their own.
Anyone else having de javu here?
Oracle HTTP Server == Apache--
Oracle JDeveloper == Borland JBuilder--
Oracle Containers for Java == Orion
The question is whether the last line will soon require a decrement operator too, or if Oracle will leave well enough alone. Given that the Oracle marketing folks have to come up with some reason to charge 10 or 100 times the current price of the product, I'm sure they will invent some lovely new "features" to differentiate their version of the product. I wish they would instead focus on providing support as an excuse to charge through the nose. Nonetheless, any change from OSE is good - that had to be one of the most abysmal product offerings in recent history.
I also hope that Oracle paid a bootload of money to Ironflare because those folks deserve it!
Agreed, and well-said...
What's that saying again - history repeats itself? What a wise man that was... goodbye m$, hello oracle. Same song, different verse.
Frankly, this whole thread sounds a bit like sour grapes...
What software company these days doesn't license technology of others?
Bottomline, smart move for Oracle. They needed the help in the J2EE world. And if anybody has the facilities to provide support, education, services, etc... They do.
I agree too. In last years, Oracle has been very "smart" in marketing stuff (and to make money) but the tech stuff... May be Larry has been very obsessed with Gates fortune...
Besides that Oracle's simply doesn't mention IronFlate on their product information pages (but only in the corporate news), I think their last move on Orion is good for the Java industry. Wall Street people (and many others capital owners) will pay more attention to small but very techical competent and (business) innovative companies like Ironflare : first class quality J2EE server for an affordable price.
But Oracle has some tech merit : Oracle Business Components for Java. IMHO I think it's a great O/R mapping framework for develop business applications. EJB 2.0 persistence model lacks the flexibility of BC4J : views objects and view links. In my experience, I often need to make more database (or entity EJB) calls to deliver the complete logical view of the entity bean to the presentation layer.
For example : my order entity has an customer_id property. This is a FK to the customer table (or entity EJB). But to show the order object I need the customer name, of course...
So with two EJBs, one for the order and the second for the customer, I need also to make at least 2 EJB calls to get all the data that make sense, correct ?
BC4j give the option to retrieve all the data into a SQL engine level - just one call. Also, with JDeveloper's support for BC4J views, writting effective 3 tier and OO business applications with Java is finally going easy !
A question : Who will pay a lot of dollars for a Oracle packaging ? The really hot component of Oracle9iAS is Orion and it can be downloaded for free. Yes the production license has a cost : just "$1500 per physical server"...
remember similar move, a few years ago, from some well known J2EE server vendor!
Can you tell me more about BC4J?
What is your organization doing with it?
oc4j is already orion--
Oracle has removed support for free databases hypersonic, postresql and sapdb. You can see what's missing in the config/database-schemas directory.
Troll bait for sure but I'm willing ...
OC4J is not orion--.
OC4J is Orion for Oracle.
So they removed mySql from the distribution - there can't be any surprises there for anyone. Wouldn't you do the same?
To the Orion distribution you failed to mention that they added their JDBC drivers for Oracle 8.1.7, provided default DataSource configurations for it, provided a first attempt at shipped documentation, and now support it.
Steal? More like buying. Since it is a better practice to investigate buy existing stable software before you would build it yourself.
Good IT practice I would say.
I can't wait for PL/SQL support to be added! ;)
J2EE is a component of Oracle9i Application Server. This means that PL/SQL is already implemented.
The PL/SQL comment was actually an attempt at sarcasm. :)
This is where Oracle are applying some clever marketing spin, because I believe the J2EE compliant app server in question is not Oracle 9i Application Server, its just the Orion App Server where the downloadable installation zip file has been renamed. If you have a look at the company package names of all the classes in the zip, they haven't even been changed to 'oracle'.
Unless you are saying that the latest version of Oracle 9i app server bundles the Orion App Server containers in with it as well. Are you saying this is the case in addition to being able to download the J2EE compliant app server on its own?
I did download Oracle9iAS 18.104.22.168 on NT, which is a monster spanning 3 CD's. I tried to install it, but I have to choose the "Enterprise Installation" option if I want to install OC4J (at least according to the prompt). This option requires 3.5 GB of disk space. As a result I did not proceed. Why bother? The OC4J download is only 10 MB and it is really the only thing you need.
Yes, Oracle did not even change the references to Orion in the jar file and the deployment descriptors. May be this is part of the license agreement?
BTW, did you all read the news from JavaOne about the exchanges between Larry Ellison and Bill Coleman? It was definitely interesting. I think if Oracle tries hard enough it may manage to grab the No. 3 spot in the app server market.
If Oracle is really interested in doing the RIGHT thing for their customers they should License BEA Weblogic as their J2EE Server. After all, 100s -if not 1000s of other companies already have partnerships with BEA (e.g. PeopleSoft, Siebel, Clarify, Broadvision, Vignette, Interwoven, WebMethods, Ariba)....
Of all people, Larry should recognize the value of selecting the market-leading solution....
Before buying Orion, the weakest part of Oracle iAS is the lacking of J2EE support.
Orion is really a great product and the speed is amazing. Why Oracle bought Orion server? There must be some reasons. They must have done lots of research on which J2EE server to buy.
I heard that they are integrating Apache (the iAS web listener) with Orion and will boundle Apache and Orion to be the core product of iAS. Once it happens, I guess iAS should have real impact in the Application Server market.
Apache and Orion are nothing new, they were options when all these other companies did their research. Why would they allofasudden become more popular at a higher price?
One might think they are buying well-researched and supported product, but its not like the support guys can take anything back to engineering and what are the chances that Orion gets purchased by IBM or that Oracle decides on a new direction "du jour" for their J2EE strategy??? (Past failures are no indication of future performance ....but I wouldn't bet my $ against it....)
I find Orion to be a much better solution most of the time and I have used both BEA and Orion extensively. Oracle already has a market-leading solution in the database arena anyway - probably having relationships with many of the companies you list. BEA may have most of the market share of the App Server world - but that doesn't make it the best software in all cases and since Larry has the market share for his products, he can choose a highly effective technical solution instead. I think it is a much better return on investment than buying BEA.
Many of the key issues in choosing a platform to standardize on relate to the benefits of minimizing internal development and support overhead and maximizing manageability, tools support, component availability/ISV support, and integration capabilities. (market leadership is a factor for all of these)
I might agree that there are isolated little apps running in small companies that would be perfect for Orion or JBoss -so my question is why would you spend more for Oracle 9ias.
I agree with you that if the business is relative small, it might be enough to just directly buy Orion rather than buying Oracle iAS suite. However, for some enterprise level, large ebusiness applications, Oracle's Apache/Orion, with possible additional Web Cache, BC4J, Discoverer, Eportal, JDeveloper, etc plus Database support should be attractive.
The point is, for Oracle, it is really a smart move to have a long time awaited J2EE container for iAS to compete in the market instead of still using the old JServ or a bad OSE.
Don't get me wrong, I too think it is great that Oracle is on the J2EE bandwagon. However, Oracle's historical modus operandi is to try to make the database the solution to every problem (e.g. Java in the Database, JMS based on the database's so-called Advanced Queueing, Oracle Portal/webdb, Internet File System, etc.) This approach only leads to a jumble of application interdependencies, slow performance and ultimately a big database bottleneck. This Oracle-centric approach has left the database as Oracle's only best-of-breed development-related product.
So, this offering by Oracle is well-marketed attempt at creating a Trojan horse to pave the way(or conceal the inadequacy) for the rest of Oracle's sub-standard development products.
(Remember the only difference between Larry Ellison and Bill Gates
......Bill doesn't want to be Larry!)
Previously, Oracle was on a database-oriented track. They tried to put everything inside the database. Some of them are right. Some of them are wrong. Some are wrong market-wise, but right technology-wise. Some are wrong both market-wise and technology-wise. :)
The existing Oracle DB JVM architecture is too heavy weight for most of J2EE web-based applications.
If you understand the architecture of database (e.g. listener / dispatcher / process memory space vs global memory space). The database architecture is targeted for a heavy-weight operation (relative to vanilla Java operations) which requires a very high level of stability. (Remember, it's not that difficult to crash a SUN's JDK JVM process. How often can you crash the whole Oracle DB completely?)
So, the Oracle's JVM inside DB is focused at stability. Oracle's session based JVM architecture avoid problem in one Java session from spreading to other sessions. This stability comes with a price. In order to maintain this session-based semantics, JVM needs to do a lot session data migration operation and those operation are expensive within database architecture. Therefore, you will get a lower performance compared with a vanilla JVM env.
The old J2EE stack running within the database basically needs to live with the same database JVM heavy weight architecture.
But, most of J2EE web applications need light weight operations and fast request-response time.
The analogy is:
the old Oracle DB J2EE stack is like an 18-wheel truck:
heavy weight, requires a lot of resources, difficult to drive, but stable.
the new Oracle J2EE stack (licensed from Orion) is like a team of small pickups:
light weight, requires fewer resources, much easy to drive (if one pickup broke down - e.g. a JVM crash, another pickup can act as a backup to deliver the same operation immediately).
The old architecture fits more into a BIG SMP box which can do a long Java Procedure call - which may last for seconds / minutes to complete - maybe batch processing - similar to your PL/SQL procedure (you cannot afford to rerun the same procedure because it is too costly to rollback and rerun).
The new architecture fits more into a farm of small servers - which matches the most of website deployment situation. The J2EE web-based operation is light weight - the frequency of method calls per object is higher and the duration of each call is short - ranging from milliseconds to seconds.
Regarding to Eric Ma's viewpoint, I agree with him with quite a few places. The old J2EE stack within the database is difficult to use, slow, not enough scalibiity. However, I do think that the old stack can give you a good stability (after you set it up succussfully)
Basically, this is a major change of Oracle's Java direction. It will give some real competition to BEA and IBM - giving people more choices, more real choices.
Personally, I really don't think IBM's websphere is that great. When I checked on their website few weeks ago, they still don't their own JSP engine. IBM just asked people to patch their websphere with Apache/Tomcat's jsp engine. They have their market share just because they bundle with their s/w with their IBM h/w. And, if you can look at their white paper document, it is just a bunch of buzz terms without real technology innovation. They partnered with M$ to push their SOAP standard.
I think, the competition about J2EE would be just between BEA and Oracle very soon, if Oracle continue to play its cards right.
Oracle finally understand the needs of most of developers. But, in order for them to catch with competitors, they must license some technology. They picked Orion, which is a smart choice. Whether Oracle can improve the new J2EE stack or just do a piggy back ride ... we will see. Time will tell us the answer very soon.
(BTW, Oracle JSP runs also on JDK stack. As far as I know, Oracle will enable OJSP to run on the new J2EE stack very soon also. And, I use JSP heavily.)
Don't leave out Pramati....these guys are the first ones to release EJB2.0 - before BEA and others..
There is some tough competition ahead for the biggies here!!
Fantastic products from Pramati
I have read about them too...but am a bit sceptical...J2EE Server at $8000...?Am not too sure...
I think it was a very smart move. Orion is well written in terms of optimization. It has a small footprint and is very resource conscious. The developers of Orion are top notch and they make themselves readily available on the #java channel in the IRC chat network. Orion always did lack support but now with Oracle it will have plenty.
BEA Is a fully J2EE compliant solution.The sheer strength of the number of ISV's and other tool vendors building on BEA makes it very attractive compared to any other solution
Uhhhh, I hear alot of whining that Oracle bought Orion rather then building it themselves. They only bought Orion, people say, after failing to write their own app server.
Sounds a bit like a large middleware company, named BEA who was floundering with their own app server until they acquired a small company who had an all java app server named WebLogic.
Short memories in this industry.
Competition in the app server space is always good. At least Oracle has finally admit that its own app server core technology, after years of marketing hype, has been just that, and has gone for the buyout approach.
However, unlike BEA's buyout of Weblogic a few years ago, Orion is a open source server with a small market share, which then begs the following questions
1) Why should I pay $10,000 (or whatever the 9i license fee) is for something I can get for free? Is Oracle support really THAT good? :-)
2) Weblogic, when it was a standalone company, had already a large marketshare, hence was production-tested. Orion, which theoretically performs better and is more stable than WL, does not have the marketshare or penetration to back up these numbers in large scale production environments. Can it do what it purports in the real world?
Also, at JavaOne a lot of good things were said about IBM's WS 4.0. I will definitely give them the benefit of the doubt and download an eval when it comes out. And finally, if Iona spend just a fraction of its resource in developing its app server as it did at JavaOne for its lavish party and contest promotion, then there's high hopes for this product as well! :-)
Both Oracle's OC4J and Orion are NOT free for server deployment. And, they are not open-source.
But they are free to download and free for development, while Weblogic charges for development licen$e.
For a company who has a lot of developers, it makes a huge difference. :)
Those are the companies that usually pay developers low salaries and work them 12 hr days and every weekend without overtime.
The market leading is not same as the technical leading. I believe Oracle made the correct choice. As a user of both weblogic and orion. My opinion is that weblogic has better marketing strategy. Instead orion has stronger technical strength. I believe orion also has good future.
I don't think that you may compare Oracle to Microsoft in that case.
When Oracle licenses Technologies such as Apache or Orion, they commit to the use of open standards like J2EE.
Microsoft instead doesn't care about open standards and always tries to push it's own proprietary technologies.
"Oracle quitely [sic] switched focus"?
Have you look at the Oracle website recently? The JavaOne presentation? I'd say their introduction of Orion/OC4J is anything BUT low key.
It's a loud shot across the bows of Weblogic and IBM, we're coming.
Well, there was definitely not a lot of publicity, until Larry Ellison raised the bar yesterday at JavaOne. I guess he just could not pass the pulic relations golden opportunity, when Bill Coleman was the keynote speaker before him.
Let's take a look at how the events unfolded:
OC4J was first made available at the Oracle Technet site's Oracle9iAS download page on 6/2/2001, and the headline about its availability did not appear on the homepage until 3 days later. So they didn't publicize this for several days. This was in stark contrast to last June, when Ellison pre-announced Oracle8iAS 3 weeks before it was rolled out. That's why I think the whole thing was "low-key", because it was so unlike vintage Ellison. Nobody really knew it was coming, unless you are a loyal Oracle customer in which case you might heard some rumors from their sales persons.
Also no where in any of the official Oracle materials can I find reference to the fact that OC4J is licensed from Orion. They just cannot openly admit their had failed terribly trying to offer a J2EE-compliant app server until now. If one does not track the history of iAS, (s)he may think all of a sudden Oracle came up with something better than WebLogic and WebSphere, all by itself. If Bill Coleman had not pointed out those earlier failures by Oracle, people outside the industry might think Ellison and Co. are the genius again. Obviously they don't want to dwell on their failures, they want to focus the public's attention on OC4J. Initially they did this quitely, but now they are really pushing it.
Having said that, I am in no way trying to accuse Oracle. On the contrary, I think Ellison is brilliant in recognizing he was losing out on the app server war and decided to look outside Oracle to find something that can help him out. Now he has found that magic bullet, he is waging a full scale offensive. The battle line is now drawn. Although this will not threaten BEA and IBM in the short term (unless Ellison bundles iAS with Oracle9i database and offers a two-for-one special, which can be a deadly weapon), if Sun and HP don't respond fast enough, iPlanet and BlueStone will be left behind in the dust.
I think the first 'public' launch was yesterday at JavaOne, hence the low publicity before hand - they didn't want to reduce the impact.
The reason there was no announcement months beforehand is precisely because they didn't have the software months before hand. The deal has been completed in the last 14 days (I'm still amazed by the speed they put together their whole JavaOne shebang).
Oracle says in their press release, and on their boards that the technology is from IronFlare / Orion. But I agree that they are trying to play down the fact it came from elsewhere, who wouldn't?
It's still not a low-key launch IMHO. The initial quietness was to tie up loose ends before JavaOne.
I just wonder if Oracle will change anything in Orion? Because I know Orion is good, but its not J2EE certified... difficult to sell ;-) However, Oracle iAS will be certified, thus Orion will be certified "indirectly"... if they didn't change anything.
Ah yes, and to IBM WAS: Its anything else but great: Practically it is both the slowest and most difficult to use appserver out there, not even compliant. But it is well integrated with IBMs products and you know "No manager ever got fired for choosing IBM"... well thos who chose TeamConnection maybe ;-)
And concerning BEA: Their Appserver is also anything else but great... 5.1 is lacking and slow, 6.0 is slow and unstable (I would not dare deploying anything on it in the "real world")... and both versions are easy to crash... too easy.
Sybase's server is great, but a bit lacking in usability.
I'll stick with Borland AS for the "Enterprise" and jBoss for those who want to have a low cost system ;-) But only until Orion gets certified, then I'd prefer it over jBoss... because a certified server is lots easier to sell.
Bottom line: Anyone knows whether they changed something?
kind regards (and sorry for the many appserver comments ;-)
As far as running logic in the database goes -it only sort of makes sense, like you said, in batch mode. IMHO batch processing will soon only be mentioned on museum tours. The need for batch processing(which is a hack technique anyway) is completely removed with a solid J2EE application design. Designing a system as an asynchronous, near-realtime system allows for better scalability (database is not a single point of failure/scalability limiter), better response time (application can acknowledge transaction and continue processing asynchronously), such a system is easier to test and debug (because it naturally allows you to check each step in a single transaction), and it allows the business to make decisions faster (you don't have to wait until "the batch runs" to find out what has happened -do you like waiting till the end of the month to find out how many minutes you've used on your cellphone?)
You may also notice that a side effect of this paradigm shift is smaller transactions -not "the query/update from
#%*#" transactions that have to run in the database to finish by morning.
Oracle has definitely not made this shift yet. Look at AQ and their Concurrent Manager architecture in their apps. They probably won't for a while because Batch mode means a bigger databa$e.
I beg to differ on batch processing. Things like doing full general ledger processing (including reconciliations etc.), risk analysis etc. (i.e. the normal end-of-day processing at a bank) require a reasonable amount of computing power (and db accesses) with no human interaction. Yes, you could do some of the processing in real time, but it would cost you response time (sometimes moving from under 1s to 10-15 sec) so is it worthwile? During night the capacity is unused, so it makes sense to use it instead of buying bigger box to catter for higher utilization during the normal business hours. This is unlikely to disappear anytime soon, as with more computing power you can do more complicated stuff (some of the risk/market analysis algorithms are so complex that running them now requires days for relatively small samples)
You know, enterprise computing ain't the same as web computing (a fact that Sun missed with the original J2EE, otherwise how could they leave out things like (mandatory) messaging support? Scheduling is something that they haven't even mentioned anywhere, yet most of the AS support it in some way or another - so why there's no J2EE standard?).
What does it do for most companies that it can put something on the web? Not much (still. Could change in future). Being able to streamline they internal workflow/processes can make a lot of difference for anyone.
Clearly you seemed biased towards your options.
I really doubt your experience with WebLogic. I know many sites running with zero downtime for Months using wls5.1.
Also I don't think wls5.1 is slow. With what servers are you comaring when u said wls5.1 is slow? What comparisons have you done?
Regarding wls6.0, did you use the latest service packs? Problably wls6.0 is one of the most stable products.
You can't through out some comments just bacause you can.
I do agree with your comments on WAS Based on my own experience.
Well, you can doubt my experience... but still, its my opinion, and I'll give some reasons:
Regarding WLS 5.1 I meant it is easy to crash during development... it often says something like "Listener no more listening", you'll have to restart... I think it is sufficiently stable for a deployment if your project doesn't have any bugs. WLS 5.1 is slow compared to BAS (Borland has a whitepaper on this, but they do not give it away anymore, due to WebLogic license)... we did a project on WLA 5.1 and deployed it "just to see" on BAS... and it was lots faster (the clients could even see this), especially with EB operations. I also cannot elaborate on this because we are under NDA with the customer and BEA (due to their license, again).
I don't have enough experience regarding Orion, but I think it is faster than WebLogic , as Oracle seems to be pushing it heavily. But I wouldn't bet on this.
WLS 6.0: No, not really, because I tried using it when it came out. But I guess the bugs in the console (which are due to the web interface) still exist... and I am quite sure it is still unstable... BEA themselves said this if I remember correctly. And they will soon be pushing 6.1 as I see it now.
But the most obvious point is: Why should I use BEA if there are so much better solutions?
You see my comment come from some experience however.
Regarding WAS I must say I don't have that much experience with it either... but what I saw and tried was enough. I'll wait for 4.0.
I tried to apply for the $1,000,000 offer for Oracle and their app server last year, and was told that the offer did not include Orion.
They knew it was faster.
Oracle is still a little behind, since they only recommend their http sever (Apache) as a front end for Orion. This isn't necessary. The Orion only solution for web pages and appserver combined is still faster than using Apache/Orion.
Bundling the e-commerce and portal solutions with orion and oracle is a BEA killer. Orion has always been j2ee compliant and used xml deployment descriptors. The documentation already available on the otn site for orion is much better than at orion's site, and it is reassuring that you can call somebody if something in orion doesn't work. This is great news.
For those of you just trying Orion, you will notice that it is much faster at auto-deployment, there is no intermediate step to compile your ejb's (as has been the case with BEA until their latest version), and if you are in deveoper mode, your servelet and jsp changes are automatically redeployed as you change them.
BEA killer? Why? Weblogic has led the Java middleware space for over 5 years now and has single-handedly made J2EE as popular as it is (Before that, BEA Tuxedo led in the market for middleware). Imagine what type of reputation J2EE would have if everyone had been forced to use Oracle or IBM's products. Weblogic also has many more features, optimizations, and 1000s more successfull implementations than other products -even orion can't touch the ease of use, manageability, clustering options, and EJB standards support. And the new Weblogic Integration Server stands to make the J2EE connector architecture equally as popular for EAI and B2B apps.
If Oracle/Orion are so brilliant were have they been for the last 3 years? Have they been driving the number of J2EE deployments, the number of successfull projects, the adoption of J2EE as a platform for enterprise computing? In short, No.
Oracle has been spending zillions on magazine ads trying to sell their "Application Suite" of packages (built using proprietary technology and a client/server design) that are so hard to integrate that they have to be "pre-integrated" and they haven't _EVEN_STARTED_ moving them in the direction of J2EE.
This demonstrates the level of commitment to J2EE that Oracle has. Do you, as a J2EE developer, want to go code PL/SQL for the rest of your life to customize and integrate apps using out-dated technology and an obsolete design???
From my perspective, as long as a product like Weblogic is available, no company is going to bet their enterprise platform on two guys from sweden(Orion)....whether Larry supports it or not.
I think you should remember that 1000's of installations does not always mean that is the best technology and neither does it always imply that the products is the technology leader. Agreed that BEA has done a very good job in the app server space. orion is a greta product too, being a cheaper server does not mean a low quailty product.
There are companies like Pramati, Persisitence which have been in this space for so long now. Do you think their products are inferior. Do you think Weblogic's clustering solution is the best. Try out Secant's solution, try out Pramati . they are all on par with Weblogic and may be much better in some front.
I think you should reflect on dimensions like what marketing and Business acumen can do to technology. being a linux fan i can say that there are more windows users than Linux, but that does not make them guys on better technology.
Your final statement is very short sighted, BEA can never always stay the world beater....reaching the top is not as tough as staying their at the top.. hope u got me
Matt - while bea has a lot of marketing clout and the ear of a lot of companies (no pun intended) having used both of the application servers (orion and weblogic - haven't tried the oracle version) I can certainly state that the "two guys from Sweeden" will easily give bea a run for their money. Big companies don't mean the best companies. If corporations are truly looking for the best solution, rather than the best name, then they will at least give Oracle a look-see. BEA is a fine product, but it is only one product. No such thing as a one size fits all solution - that's the idea of competition. Other servers offer their strengths. Competition will make better products.
Weblogic has led the Java middleware space for over 5 years now and has single-handedly made J2EE as popular as it is (Before that, BEA Tuxedo led in the market for middleware).
That's right, but...
1) WebLogic was developed by company called WebXpress (and it was called Tengah). BEA bought the WebLogic, when it became popular and (of course) made it more popular.
2) Tuxedo is much more older the BEA itself (it was developed in Bell Labs, then bought by Novell, then... - I don't remember the whole story, but it's rather long and complex)
3) The number of technologies bought or licensed by BEA is rather large (BEA licenses rule-engine from ILOG for they personalization server, bought Theory Center suite of components, licensed engine for WebLogic Process Integrator etc.)
Tuxedo was BEA flagship, WebLogic is BEA flagship. It's quite normal for middleware market to buy new product or companies or license new technology from the others. Oracle move was clever - both from marketing and technology point of view. If the quys that built Orion were really so good, they got their chance. Oracle marketing, support and sales machine can make their product more popular. More competition on J2EE market means better solutions.
You say that if Oracle bundles its portal solution with Oracle DB and Orion it would be a BEA killer.
Perhaps you should take a look at Oracles portal product closer. You will see that its isnt pure Java and needs to be rewritten or replaced (Orion styleee).
Orion is some quasi-shareware written by some college students in Europe as I understand it. I don't know why Oracle is using it. Maybe they took their engineers and cleaned up all the bugs and implemented some scalability.
Orion is some quasi-shareware written by some college >>students in Europe as I understand it. I >>don't know why Oracle is using it. Maybe they took their >>engineers and cleaned up all the bugs and implemented some >>scalability.
Apparently you haven't used the product! Instead of listening to what other uninformed people tell you, go try it your self.
No one familiar with OEM deals could consider this "stealing." Orion gets: (a) cash and (b) Oracle's assistance with db-related issues, such as implementation of abstract schemas (EJB 2.0), complex data caching, and a full DTP.
Similar deals abound throughout the industry, as Merant drivers are embedded and re-branded, the Resin and JRun servlet engines are embedded in other app servers, SonicMQ is embedded as the JMS provider in many app servers, etc. Targeting the OEM/ISV market instead of end users (with the tremendously expensive service arm those users require) is a perfectly valid strategy for lightweight J2EE servers and containers.
No one familiar with OEM deals could consider this "stealing."
I can't believe a few of you guys actually took my words literally; obviously I didn't mean they were "stealing" anything. What I said is that they acquired orion the easy way - like micro$oft did with everything they have, which is by simply buying it. They are the m$ of the mellenium; they either smash you or buy you.
Same song, different verse. Their database isn't sql compliant, you can't even rename a column, which is a standard sql statement, outer joins (=*) aren't sql compliant and are intrinsic, and their fees are extraordinarily out of line with what you get. Jdeveloper? What a joke.
Majority of RDBMS are not SQL compliant and do not complettely folowMajority of RDBMSes are not SQL compliant and do not follow relational algebra completely. So what? They are still usable and serve busyness needs.
Oracle, Microsoft and others are not in business of creating superior technologies and releasing new tools they are in business of making money. So everything goes!
Accusing someone of buying competition is like shaming company for exercising sound business sense.
My point is that m$ was always lifting their nose at standards bodies on the idea that they were too big to conform and they tried to make their own standards. That isn't good for the industry.
Are you saying it's perfectly ok to not conform to a standard?
No it is not good for developers and technology when companies do not conform to standards. But it is the reality that they do not. And it is not just Microsoft that has this fallacy?
Plus, companies that conform to standards still try to provide proprietary extensions of them. This is just a part of fundamental problem concerning innovation and how it is done today.
While 10 MB is very impressive for a 'complete' app server I'm wondering where the object caching library files are. They were not in the standard install and there doesn't seem to be any place you can download it.
I'd like to hear early feedback from people that are currently evaluating it and see what issues there are...
Disclaimer: I work for Oracle in a pre-sales position. These opinions are my own and not those of Oracle.
Hmm, didn't BEA buy WebLogic AND Tuxedo. When Sybase needed a development tool they bought Powerbuilder. IBM just bought Informix. Lots of PeopleSoft's CRM solution was purchased. Siebel has bought a bunch of companies to round out their solution. Happens ALL the time in this industry. Frankly, I'm glad Oracle bought or licensed (I don't know the terms of the deal) such a quality product as Orion. It was the only way to be competitive in the J2EE Application Server Market.
.It was the only way to be competitive in the J2EE Application Server Market.
Yes, you're right - they failed to make one on thier own, so yes, they did have to buy one in order to stay competitive.
If you can't beat 'em, buy 'em.
I think that is a good deal for Oracle to sell Orion as part of their solutions.
But the real questions are:
- What is wrong with Oracle development?
- Why they couldn't develop this if they had much more resource, people, knowledge?
- Can the clients trust Oracle technologies when they can't trust their own?
Personally I think that there must be something wrong with Oracle development philosofy. The only thing that keep they alive is the database. They should try to be more productive.
Another questions are:
- Orion will get the right payment for selling they work?
- Is Oracle fooling Orion into a deal where the biggest winner is Oracle?
- What was the deal?
- Why Oracle doesn't say that the maker is Orion? I haven't seen Ellison saying this. He says that was Oracle the maker.
If Oracle is paying right price, then I don't see a problem. If Orion can develop something and went to Oracle and got a lot of money Orion must be very happy.
Orion can't compete with Oracle marketing and fame, but if Orion can get the money and the recognition for its work and creativity is fine.
So what a lot of the converation on this thread boils down to is, in what stage of development is the application server market?
Clearly BEA and IBM have market share measured by both revenue and installed base. Also, most anybody that takes a objective look would have to agree that the market is relatively mature by high technology standards. Taken in combination this means that either BEA or IBM will end up being the gorilla, as these markets will always pick one top banana. I have my own opinion on which of the two will come out on top. That's not, however, the point here.
The point is that any and all of the other players will be relegated to chimpanzee status over time. Sure Oracle has the financial clout and marketing muscle to make a noisy entrance to the party. Sure Orion is a good technology that has its merits. The Wankel engine was too.
But does anybody really think Oracle they can buy their way into the market? What do they expect the net new growth in installations to be? How big of an installed base does Orion have today that can be leveraged? Companies have been buying application server technology for how many years now? To be sure, room exists for new license sales, but BEA and IBM will continue add new customers too.
Maybe Oracle's bundling of Orion is the next best thing compared to slice bread. I'm willing to bet, however, that too much market momentum exists for Oracle to dominate the game now. Time will tell.
Surprisingly nobody's pointed out the irony of Oracle's strategy.
Does anybody remember how Oracle vaulted itself to market leading status in the 80's? They did it by making PORTABILITY an issue. Oracle runs on Vax VMS. Oracle runs on Unix System V. Etc. Portability implies choice. J2EE implies choice.
Somehow choice looses its meaning when my application server and database management system are joined together at the hip.
As an independent consultant, there are large clients that I have who are very interested in the new software from Oracle - they WANT more competition than BEA and IBM. Perhaps Oracle/Orion won't have the footprint that BEA or IBM have but that remains to be seen.
You mean like the choices we have in the RDBMS market? Who is the "standard" in that market?
BTW I'm not arguing that its "right" for technology markets behave in this fashion. They simply do. Look at the evidence. Look at the behavior over time.
Who owns the desktop? Who owns the mainframe market? We aren't talking about rational, objective, mass decision-making.
Same disclaimer as last time: I work for Oracle. These are my opinions and not those of Oracle Corporation.
Oracle's strategy was to support J2EE in the database via JServer. Turns out that it didn't work very well for J2EE. However JServer (or Oracle JVM if you like) does work very well for Java Stored Procedures in the context of database sessions and database transactions. So Oracle took the most direct route to a competitive J2EE offering.
Somebody will probably end up being the 800 lb gorilla of the app server market - BEA or IBM. Despite Oracle's lackluster mid tier offerings in the past I do believe that OC4J/Orion will open many doors for Oracle. Heck, even if only existing Oracle DB customers kick the tires, that's still a lot of potential clients.
Looking at Sun's site it appears that there are around 18 or 20 J2EE servers on the market. This market will probably consolidate down to three or so big players, and a similar number of niche vendors. Maybe it's my database bias showing but the J2EE market looks a lot like the RDBMS market of 1992/1993 - Oracle, Sybase, Informix, Ingres, DB/2, and smaller players. I think the market will eventually look like the RDBMS market looks today - three big players (Oracle, IBM, Microsoft), some second tier players (Informix, Sybase), and then maybe the open source offerings. Will Oracle be in the top tier? Only time will tell.
Just my worthless two cents.
Personally, I think the database vendor will have a big advantage in the app server market. There are constant complaints about performance especially in the area of persistence. I would think the database vendor with control over the DBMS would be able to engineer optimizations that would be out of reach for the others.
Personally, I think the database vendor will have a big advantage in the app server market. There are constant complaints about performance especially in the area of persistence. I would think the database vendor with control over the DBMS would be able to engineer optimizations that would be out of reach for the others.
In my (humble) opinion the main aim of application server is to hold application logic. That's what are they built for. In typical application you have presentation, application and data layers. In the enterprise applications they should be separate and independent. In large companies you have different databases (Oracle, Sybase, Informix, IBM DB2), legacy applications etc. The performance is of the essence, but you must also consider integration. In complex projects, your "persistent" store may not be Oracle, may not be even relational database. It may be the mainframe with some CICS application.
Talking about application servers is a discussion about middleware. And the middleware is BEA's and IBM's domain.
That sounds great but if the technology is desirable but yet broken, concessions are made. If no other product is able to satisify some performance requirements but one that
only works on some of the most widely used platforms then it will get market share. Performance is a very hard requirement in this domain and having kernel constructs such as entity beans fail performance is a larger problem than platform independence to many customers.
Perhaps the application server market is mature - MAYBE - but the packaging of them is far from that. BEA is doing a good job with horizontal integration of the enterprise and I think that is the future direction - plus the potential of J2EE has just barely been hinted at - its an evolving architecture involving a growing number of technologies - but (back to BEA) other companies and their products, such as Oracle or Sybase with EA Server (as an example), or a host of others can absolutely package competitive products. I truly hope they do. Market share is a funny thing and so is momentum - it can shift. Having it now doesn't mean you'll have it tomorrow. My clients large, medium, and small all want a choice in the market place. If (and it certainly is a big IF) Oracle markets, supports, develops and packages everything correctly - they can at least make a solid run at taking a bite out of BEA and IBM's market share.
I think a lot of the contention raised on this thread can be boiled down to the following:
Oracle has failed at bringing an appserver to the market (they have failed twice and this 3rd attempt is weak -why buy Oracle when Orion is $1500???)
Every consultant/developer cannot know how to effectively use and administer _EVERY_ app server that comes to market -regardless of its technical merits. This means that for Job security they will try to know at least 2, maybe 3.
Most career-aware technologists will choose based on market leadership. (Just as most companies do). I shouldn't need to point it out but this behavior creates a vicious (or virtuous) circle where the market consolidates around 1-2 products (i.e BEA WebLogic).
Therefore, Pramati, persistence, ea server, orion, yadayadayada will never "play" at the enterprise level.
It seems that all we get from BEA boils down to this:
No one can, or will, compete with BEA.
I can tell you that there will probably a good deal of buyers for it SIMPLY because it is oracle. I mean, look at the facts; one of the most horrible abominations of an app server ever (oracle app server) STILL is out there, and that's because it's oracle.
SO, if oracle actually puts something out that is technically sounds - whether they stole it or not - it will probably sell. It'll be a while before they get momentum like bea or websphere though.
Oh, almost forgot... DOWN WITH ORACLE!! :)
No momentum over the next few months? Dunno about that statement. They're a pretty big company that might have something to actually do the job they tout it to do for once, and if it does, then you don't think they'll be able to sell it?
If OC4J really is fast, is standards focussed, and Oracle does a good job of keeping it that way, then isn't that a good thing?
I'm interested to hear what you think engineers and IT management care about? Simplicity, standards adherence, openness, performance? Are you saying that people are that tightly bound to an application server platform from a specific vendor such as BEA/IBM that they want to have only two vendors in the market to select from? I was under the impression that J2EE is the technology that is being championed and we're all pretty happy to get vendors' best efforts at implementations.
Personally, I don't see what's wrong with a little competition in the market place.
And did Oracle steal it? Ever hear of the word tengah, or a company called Novell? I'm sure the fellas from Sweden were more than happy to participate - they must have gotten some cash from the deal and can now keep developing without needing to worry too much about where the next meal is coming from. If they're full time on development, with heavy wallets to boot, who knows how good Orion/OC4J will become?
Why don't you pull Oracle's latest stuff down, and try and run some of your J2EE apps with it. Then write and tell us what you think based on some usage of it.
"If OC4J really is fast, is standards focussed, and Oracle does a good job of keeping it that way, then isn't that a good thing? "
It will probably be that way for the initial release (i.e., with Orion in its pure form), but after oracle starts mod'ing it, it won't. cya
I haven't heard any of the details of the licensing agreement they had with the Orion team -but- if they were smart enough to license it in the first place, perhaps they'll recognise that and maintain that relationship to continue to base OC4J on Orion?
When you look at it, why wouldn't they? It'd be a good move I reckon. Let the Orion developers who produced a pretty sweet product with apparently very little funding focus on the core development and continue to ship that as the J2EE server piece of 9ias. That'd keep it fast, lightweight and up to date.
Excerpt from a recent company research report from Deutsche Banc Alex. Brown:
Oracle is placing high expectations on the latest version of its 9iAS application server and is going after IBM and BEA Systems "guns a-blazing". Management sees 9iAS contributing as much as 20% of total server license revenue in 1-2 years(currently 95% database). While we do not expect Oracle to leapfrog competitors such as BEA, IBM's WebSphere or Sun's iPlanet over the near term, we believe 9iAS could be a wildcard in the rapidly consolidating app server market. With the release of the latest version of 9iAS, Oracle has made significant improvements in its technology base and reference accounts. The company retired its older (and performance-plagued) Java engine in favor of a third-party J2EE-certified EJB container from Orion. While far from proven technology, we believe the Orion product (created by two developers aged 22 and 24 years
old in an apartment in Sweden) has cultivated a small but loyal audience of developers. In addition to this new Java engine, 9iAS also includes capabilities for portals,
workflow, clustering, caching, application integration, reporting and other features.
On the customer side, Oracle announced 70+ reference accounts for 9iAS including live customers Alcoa, Boeing, National Institute of Health, Raytheon, Nortel, US Army,
Princeton University, Sprint PCS, Citicorp, Barclays Bank, Siemens, CalPERS, Digital River and others. Additional companies that have licensed 9iAS include British Telecom, General Motors, Healthsouth, Lucent and US Postal Service. Competitive wins over BEA have included Boeing, National Institute of Health and CalPERS. In total, Oracle announced
that 15 customers have purchased over $1M in 9iAS licenses (including Sprint PCS, Citicorp, GM and Lucent), including seven over $2M and one over $3M. The number of seven-figure
license deals is even more surprising given that market leader BEA did not close any $1M+ WebLogic deals in its most recent quarter (F1Q) and closed only a handful in its
F4Q. Oracle also indicated that 350+ software partners are porting their applications onto the 9iAS platform(including Documentum which ported from BEA in less than one hour), and
over 100 systems integrators are ramping up on the platform.
Are we seeing an immediate impact of OC4J already?
Are we seeing an immediate impact of OC4J already?
I agree with all those who say Orion is a good aplication server. I have used it my self and i found it supports EJB1.1 spec like xml based deployment descriptors etc long before weblogic could have it so they have a better experience in the market than welblogic server in EJB1.1 apsect (Orion app server is similar to weblogics 6.0 + versions)
It is now on Oracle peoples hand how they fit Orion into the market.
hope they bring it in a better and faster way in the market so that people can have a taste of a good quality J2EE application server
Wow, you're pretty hard on bea.
They're a bigger company and so they move a little slower; give a little slack for that. Also, try comparing Orion's support (which there is none) again bea; if you have some weird app-related server issue giving a cryptic error, you can get on the horn with bea and they'll help you almost immediately. You'd have to learn swedish to get that from orion. cya
How much experience do you have working with Orion? You are right - Orion will not be a BEA - but so what? I hope BEA continues on its merry little path - but having used both products and if a company has experienced J2EE programmers I have no hesitation in recommending Orion (and OCJ4). There's no reason to use a product because they have more programmers or a bigger marketshare or they have someone who will pick up the phone when you call - if my customers pay BEA prices, they expect that. At the end of the day, the product has to perform the way the company wants it to. If its BEA, fine - but Orion can play with BEA any day in the app server world. Are there things that BEA can do that Orion can't? Sure. But not all companies need everything BEA offers. And not just small or mid-size companies - but the largest as well. In a very large telecommunications company that is one of my clients, they have EA Server and also Oracle 9iAS/OCJ4 (they have some J2EE guys that use Orion so it was a no-brainer to make that decision). Not a WebLogic deployment in site. And EA Server is a great app server, too! BEA may be the biggest on the block, but it will never be the only one on the block.
My comment wasn't against orion, but rather in defense of cricism against bea not supporting the very latest standards yet. That takes time; I would rather a company test thoroughly and make sure the product is stable and just throw some code in and roll it out the door.
And to answer your question, heck yes support is crucial. There isn't any with orion (yet). If money was a very big issue I'd recommend something like orion, otherwise I'd recommend wls or websphere. cya
You are absolutely right. The victim of the Micro$oft is becoming the next Micro$$$$oft......
But frankly speaking, if you want to become a large and I do mean LARGE software company you'll need to offer products in all market segments and if you're even vaguely profit-motivated(capitalistic if u please - unlike Linux Torvalds)you'll eventually end up doing what Gates did or what Ellison is currently doing. Inspite of all the anti-Micorsoft blather that Ellison has been spewing, deep down even he knows that the only way to become the 'THE SOFTWARE COMPANY' is ............the Micro$$$$oft way.
Other wise can anyone explain where an Oracle Portal or an Oracle 9i AS fits into a hardcore pure-play database vendor's core competency...? And pls..no excuses like synergy, parallel market development ....etc.........!!!!!!
The problem is we are afraid of taking risk to evaluate new and stable technologies and use it in production.We want to always play safe and follow the industry so that we wont be blamed.
Unless people (for eg Google who proved that linux can be used for high scale applications) demostrate , we will never touch products like orion.
I still wonder why such a big company with lot of geeks cannot produce a appserver that is robust?
I have not tried orion..but comparing WLS, iPlanet and WebSphere..my vote is to WLS.