News: Interview with Thomas Kurian on Oracle and the AppServer Market
CNET news.com is hosting an interview with Thomas Kurian, Senior Vice President at Oracle and the man credited for the strategy and growing marketshare of Oracle application server over the last couple of years. In the interview, Kurian discusses growing appserver market, the need for appservers, Web Services, and how Oracle is going to take market share from BEA and IBM.
- Posted by: Floyd Marinescu
- Posted on: May 23 2002 18:40 EDT
Read Can app servers revive Oracle?.
- Interview with Thomas Kurian on Oracle and the AppServer Market by andi kusnadi on May 23 2002 23:35 EDT
- Interview with Thomas Kurian on Oracle and the AppServer Market by Glen Mazza on May 24 2002 14:10 EDT
- Interview with Thomas Kurian on Oracle and the AppServer Market by Edwin Khodabakchian on May 24 2002 03:20 EDT
- Oracle´s future by Hans van Buuren on May 24 2002 04:21 EDT
- Interview with Thomas Kurian on Oracle and the AppServer Market by Greg Turner on May 24 2002 17:30 EDT
- Interview with Thomas Kurian on Oracle and the AppServer Market by Cary Bloom on May 25 2002 12:54 EDT
- Interview with Thomas Kurian on Oracle and the AppServer Market by Kevin Leong on May 25 2002 20:45 EDT
- Interview with Thomas Kurian on Oracle and the AppServer Market by Kevin Leong on May 25 2002 21:02 EDT
- Interview with Thomas Kurian on Oracle and the AppServer Market by Abhay A on May 27 2002 21:51 EDT
- Interview with Thomas Kurian on Oracle and the AppServer Market by Jon Tirsen on May 28 2002 06:58 EDT
Oracle9i Application Server Clustering (for High Availability) by Sudhakar Ramakrishnan on May 28 2002 07:25 EDT
- Oracle9i Application Server Clustering (for High Availability) by Cary Bloom on May 29 2002 01:06 EDT
- Oracle9i Application Server Clustering (for High Availability) by Sudhakar Ramakrishnan on May 28 2002 07:25 EDT
I think App server market share will be dominate with Oracle 9iAs, because they have design pattern build from real world application you can go to link to
and they had been good made some xml transformation to their own database, because of that oracle will never be lose.
Posted By: andi kusnadi on May 23, 2002 in response to >this message.
Excellent example for someone learning J2EE--thanks for the link.
Oracle is indeed starting to have a good core application server and the 9i database rocks. But the challenge here is that the next battle is about integration and web services: IBM has integrated Crossworld and have a very strong support for Web Services, BEA has Workshop and a maturing JMS solution. It will be interesting to see how Oracle reacts.
Oracle will always be a company to reckon with, they have huge market share in the database market. As the market for graphic development tools matured and took the web turn their old development tools went out of style even more. They are now finally really done for. And even so they are used widely. Oracle had to step into the Appserver market to keep competing. Now their product is still trailing IBM WebSphere and BEA Weblogic for openess and functionality. Still, because of the large amounts of Oracle fans (and frankly I think there are better databases available at lower costs) they are right to think it will start to make it´s way up. I expect two problems: first because of the maturity of their product clients may run into problems when deploying in large and heavily used environments. Second I expect the Oracle solution to be proprietary in some ways as all of their solutions have always been. Most clients put up with it as it is their strategy to stick with Oracle anyway (just listen to their managers) but other clients won´t like that. I cannot be sure about what this will mean to their market share. We´ll have to see... Surely there will be at least four big products competing in this market and that is good for us clients.
I surely would not advise to step into this product myself. Our company is using IBM WebSphere.
Hans van Buuren
Senior Systems Architect
Can you imagine if Microsoft had put a decent J2EE app server on the market from the beginning? They'd be further ahead than with trying to re-invent the wheel.
The .NET decision has permanently opened the door for others like BEA, Oracle etc.
just a thought :)
They'd be further ahead than with trying to re-invent
> the wheel
First of all, before I get flamed - I should admit that I'm primarily a J2EE/CORBA guy, and I was at the BorCon (Borland Conference) at LA this week where Microsoft was a sponsor (among others including J2EE vendors). There were quite a lot of sessions (very few from M$ though) on the .NET framework and I must admit I was quite impressed.
The entire platform & framework seems to have been well thought out and the tools around it seem more than capable. And they've gotten betas running on other OSes as well. I'd go as far as to say that people who casually dismiss .NET as nothing more than a M$ ploy (which it very well might be - but that's not the discussion at hand) are doing so out of plain ignorance.
Its something I'd watch out for and even learn a bit given half a chance :)
> And they've gotten betas running on other OSes as well.
I know that they have an example of the CLI implemented on FreeBSD, but that is not a "beta" nor is anyone allowed to use it commercially nor is it an attempt at a full .NET implementation. Do you have more data on these "betas"? I'm very curious; it would not be totally surprising to see support for the Mac (Microsoft's "other" platform ;-) for example.
Have a look at Mono: http://www.go-mono.com/
"The Mono Project is an open development initiative sponsored by Ximian that is working to develop an open source, Linux-based version of the Microsoft .NET development platform."
And another link - more business focused: http://www.osopinion.com/perl/story/7007.html
Yes, I'm aware of Mono and Halcyon, and I'm aware of Microsoft's FreeBSD (sub-)implementation (comparable to a Java 1.0.2 interpreter minus AWT). First, Mono can be sued by Microsoft, and Microsoft execs have stated as much (veiled threats, etc.) so who is going to bet on that? What I wanted to know is whether Microsoft itself has moved toward (or even feinted a move toward) supporting any other platforms with a commercial release.
I don't get why discussions of this nature fail to include the JBoss app server and its business model as outlined by Marc Fleury in his interview here at TheServerSide.com site. I think that because of its superior design and because of its business model, JBoss will end up being the app server to beat. Especially when app servers on all kinds of hardware web aware devices becomes the norm.
it is very easy to grow 96% in every quarter when you only have a customer or two to start with!
I attended Oracle 9iAS V2.0 Developer Workshop last week. OC4J, Oracle's lightweight EJB container, is very impressive. You do not have to bound the server to deploy new and re-deploy applications. Changing the deployment XML files will trigger OC4J to deploy the applications. This is great for development. OC4J also tightly integrate Apache. Clustering is easily accomplished *simply* by updating a couple of deployment XML files. Oracle 9iAS V2.0 Enterprise also includes Web Cache. Oracle's Web Cache is based on Edge Side Include (JSR 128).
The Developer Workshop is free for developers committed to port their EJB applications or components to OC4J.
I should add that the Developer Workshop's course materials are still based on 9iAS V1.0. Therefore, there is still a lot of Orion stuff in the course materials, which are irrelevant with V2.0. I have provided Oracle with feedback that they need to update the course materials for V2.0 and to organize a World Tour to reach J2EE community.
Kevin - where can I get more information about the Oracle9i Workshop? Also are there any online resources for the same?
Ehrm... With the risk of sounding a bit blunt: Dynamic redeploy and Apache integration has been around for many years now with other appservers.
As I was reading through this thread, I realized a few posts refered to Oracle9iAS' clustering features and felt it would be helpful to post a followup on this topic.
Oracle9i Application Server Clustering Overview
Clustering involves grouping together of independent nodes to work together as a single system. Oracle9i Application Server uniquely supports three levels of clustering in the middle-tier:
1. J2EE clusters
2. Web Server Clusters
3. Web Cache Clusters
Oracle9i Application Server enables the creation of J2EE “cluster islands” – a technology that is hardware and operating system independent – providing automatic failover of both stateful and stateless J2EE applications. Administrators can easily scale hardware by upgrading to a new low cost platform without rewriting their applications.
Oracle9iAS Web Clusters enables http processes to work in a cluster configuration to support automatic failover and efficient resource utilization.
Oracle9i Application Server enables web caches (which are used to improve performance of any web sites by storing static/dynamic pages) to be deployed in a clustered environment thereby increasing total cache capacity and application scalability and availability. This enables further cost savings as you can serve more content with the same hardware.
SOME FEATURES OF ORACLE9iAS' CLUSTER ARCHITECTURE:
1) More Performant: Oracle9iAS supports asynchronous replication. Other application servers support synchronous replication. We do in-memory lazy replication and the transport for the replication is IP Multicast. Multicast is used for group discovery, replication.
2) Configurable depth of failure (depends on island in HTTP). Other application servers have a designated secondary and thus a rapid failure of 2nd node can leave the system stranded.
3) Concept of islands in HTTP clustering (different granularity of loadbalancing and failover)
4) Programmatic mode for SFSB (Stateful Session Bean) clustering with setAttribute can be extremely performant. Other application servers don't support this.
5) Oracle9iAS uses client container for application client load balancing enabling you to take advantage of new nodes as and when they become available. Other application servers use replica aware stubs (with static binding)- with this approach you will not be able to utilize a newly available node.
6) Cache clustering (only application server vendor to provide this) - to reduce the risk of denial-of-service attacks, and improving availability of dynamic/static content, Oracle9iAS Web Cache enables multiple cache instances to work together as a single logical cache.
7) End-to-End Clustering: Oracle9iAS supports three levels of clustering on the middle-tier. In addition we support clustering at the dabase level with Oracle Real Application Clusters and Data Guard.
8) Fast Start Fault Recovery Architecture: Oracle9iAS introduces the Fast Start Fault Recovery Architecture(TM) to ensure graceful recovery of the application server in a clustered environment with no loss of service. This feature eliminates downtime in the event of a fault.
Principal J2EE Evangelist
Can you spell crapola? :-)
I've tried your clusters and it just doesn't work (repeated crashes and load-balancing never worked). So much so I promptly trashed it (should've known better) and switched to one of the well-known *real* J2EE app servers.