Discussions

News: IONA Artix: Video Tech Brief

  1. IONA Artix: Video Tech Brief (8 messages)

    John Davies, technical director and head of research at IONA, tells us about Artix uses in the real world. Artix is a SOA infrastructure suite with components for ESB, mainframe, and Java Enterprise Edition components. John tells us about a high-throughput deployment that combines with Mule ESB or potentially others open-source ESBs with Artix, and gives overall advise on how to build a high-performance, high-troughput application. When and where to use an ESB, determining when to use ORM like Hibernate, and combining heterogeneous systems are some of the topics that he covers. (Click here if you can't see the video.) John has nearly 20 years in Investment Banking and over 25 years in IT, mostly as a consultant. He has co-authored several books on Java and J2EE, was the author of Learning Tree's distributed Java course and a regular speaker on grid, Jini and JavaSpaces in the Java and banking world. Over the years John has held more than one high-profile position as Head of Technical Architect in banks such as JPMorgan. You can learn about John's creative side by visiting his photo blog Watch other Tech Briefs

    Threaded Messages (8)

  2. Re: IONA Artix: Video Tech Brief[ Go to top ]

    John I enjoined your presentation. As always you were able to integrate different idea's and products in that case, to single architecture that solves the most challenging problems in transaction processing. A combination of Artix as the data-source handler and Mule as the ESB sounds like a good alternative architecture for transaction processing. Put that ontop of GigaSpaces/Azul and you get ultimate scaling, processing. If i heard correctly you are talking on up to 10k messages/sec of complex Switft messages. If that is the case this is well beyond what most systems can handle today. Toward the end you suggest a very interesting model for overcoming the Data-Base bottleneck by storing immutable versions of the same objects instead of updating the same object over and over again. This sounds very similar to the CVS model. Can you elaborate a bit more on how this model works and how it solves the data bottleneck? Nati S. GigaSpaces
  3. Re: IONA Artix: Video Tech Brief[ Go to top ]

    Hi Nati, Having spent decades working for banks I've gotten used to solving problems with any and all available resources. To the contrary many but not all vendors often look for problems to solve with their products, this is the wrong way round. The talk I gave at JavaZone was about the solution several of our customers are using in production using Mule with IONA's Artix Data Services (formally C24's Integration Objects). The point being to demonstrate that Artix Data Services generates POJOs that will run with any ESB, container or application server; Home-Brew, OpenAdaptor, Camel, Fuse, Mule, WebLogic, Web'phere, GigaSpaces etc. When it comes to scaling into the seriously high volumes and/or complexity then GigaSpaces is an excellent platform due to the clean and efficient implementation of the master/worker pattern. Because it can use a "local" call or call by reference for tasks (Entrys, sic) it works beautifully on the Azul box. We probably over did the benchmark by using SWIFT MT to MX (ISO-20022) transformations in that we could convert the entire planet's SWIFT messages in just a few minutes on the top Azul box, we were getting well over 100,000 messages per second. In SWIFT this level of performance is pointless but it comes into its own when processing complex derivatives (i.e. from FpML). With this combination of technologies we can perform complex derivatives matching and reconciliation several orders of magnitude faster than existing systems. Getting rid of the relational database is really the key, far too much time is being spent on ORM to map objects to relational databases. Get rid of the relational database and the ORM and you save huge amounts of time, complexity and money. It wasn't CVS but SubVersion (SVN) that was the key to a simple solution. All that's needed is a system to store hierarchical data (XML) as immutable data, i.e. copy on write (COW). If you use WebDav on SVN then you've pretty much got a hierarchical storage system already, you can go back to any previous version and even upgrade versions. There's obviously more to it that this and SVN is way too slow to use for anything practical but file systems like ZFS are much closer to providing a complete and elegant solution. There are a lot of banks working hierarchical persistence at the moment, watch this space. -John-
  4. Re: IONA Artix: Video Tech Brief[ Go to top ]

    Hello John, Very interested presentation and very beautiful natural pictures on your mentioned personal blog. I'm evertime interested to hear from practical experiences and have a little question in relation to the mentioned combination of Artix Data Services and Mule. Existing in this case special advances by Mule, which has convinced your customers to favour Mule before other ESBs, how eg. the familiar Artix ESB, Apache ServiceMix (FUSE ESB), JBoss ESB, BEA Aqualogic ESB, ... ? Roland SOA Competence Network
  5. Re: IONA Artix: Video Tech Brief[ Go to top ]

    Hello John,

    Very interested presentation and very beautiful natural pictures on your mentioned personal blog.
    Thank you, the link to my blog was a bit of a surprise, it's somewhat out of date, I've been working on my web site and had ignored the blog.
    I'm evertime interested to hear from practical experiences and have a little question in relation to the mentioned combination of Artix Data Services and Mule.

    Existing in this case special advances by Mule, which has convinced your customers to favour Mule before other ESBs, how eg. the familiar Artix ESB, Apache ServiceMix (FUSE ESB), JBoss ESB, BEA Aqualogic ESB, ... ?
    As a company IONA would prefer people to be using Fuse and Camel but I think it's equally important to let the customer choose what's best for them. There are different sales models involved in commercial and open source and if we can solve a customer's problems with a combination of our commercial products with someone else's open source (or even commercial) ESB then we're more than happy to take the business. Obviously Camel and Fuse are well integrated with Artix Data Services (our old C24 product) and a single solution offers many advantages but it also works well in Mule and is production proven so we offer customers the luxury of choice. Choice and openness are more important than a closed, single source model. I've recently had a "play" with Camel and it's has some interesting features, after a bit of work with James I hope to be able to post some interesting benchmarks and examples very soon. Anyone interested in the raw Artix Data Services examples and benchmarks can find the source here. -John-
  6. Re: IONA Artix: Video Tech Brief[ Go to top ]

    I think it's very good and very fair when a customer can select between different solutions - out of the way from usually business criterions. All previously mentioned solutions are very great and (AFAIK) in a stable state for use in production environments. Maybe, sometimes has customers problems to identicate the right solution for their real business projects. What will mean: Many customers are don't know what are concretely the advances of each solution in relation to their real and personal requirements.
    I've recently had a "play" with Camel and it's has some interesting features, after a bit of work with James I hope to be able to post some interesting benchmarks and examples very soon. Anyone interested in the raw Artix Data Services examples and benchmarks can find the source here.
    I'm looking forward for this benchmarks and examples - BTW: nice family (C24 & LogicBlaze) which has getting together ... Roland
  7. Re: IONA Artix: Video Tech Brief[ Go to top ]

    I think it's very good and very fair when a customer can select between different solutions - out of the way from usually business criterions.

    All previously mentioned solutions are very great and (AFAIK) in a stable state for use in production environments.

    Maybe, sometimes has customers problems to identicate the right solution for their real business projects.

    What will mean: Many customers are don't know what are concretely the advances of each solution in relation to their real and personal requirements.


    I've recently had a "play" with Camel and it's has some interesting features, after a bit of work with James I hope to be able to post some interesting benchmarks and examples very soon. Anyone interested in the raw Artix Data Services examples and benchmarks can find the source here.


    I'm looking forward for this benchmarks and examples - BTW: nice family (C24 & LogicBlaze) which has getting together ...

    Roland
    Agreed! I think the C24 and LogicBlaze folks are a great combination along with the other great folks at IONA. Am looking forward to improved integration between the FUSE and Artix Data Services products. (I'm hacking furiously a nice demo of using Artix DS and Camel...) James Iona Open Source the Enterprise Way
  8. Re: IONA Artix: Video Tech Brief[ Go to top ]

    When it comes to scaling into the seriously high volumes and/or complexity then GigaSpaces is an excellent platform due to the clean and efficient implementation of the master/worker pattern. Because it can use a "local" call or call by reference for tasks (Entrys, sic) it works beautifully on the Azul box.
    We probably over did the benchmark by using SWIFT MT to MX (ISO-20022) transformations in that we could convert the entire planet's SWIFT messages in just a few minutes on the top Azul box, we were getting well over 100,000 messages per second. In SWIFT this level of performance is pointless but it comes into its own when processing complex derivatives (i.e. from FpML).
    John, That may be overkill for SWIFT and FIX messages, but (as you of course know) it's necessary in telecommunications, the other domain you touched on in your discussion. The ability to process that many CDRs in realtime makes some very interesting new applications possible. I'm sure we'll be seeing more of you at GigaSpaces. Patrick
  9. Re: IONA Artix: Video Tech Brief[ Go to top ]

    Hi Nati, One of the things we wanted to demonstrate was using Artix DS, Mule and GigaSpaces (we just ran out of time). I am very interested in a combination like this since you get a great combination robustness, performance and scalability plus we're seeing a few of our large customers that are interested in this approach. It's worth noting that Mule works with Volante data services in much the same way and Mule also works with technologies like GridGain. So what John and I demonstrated was an alternate pattern for doing Transaction Processing, where each piece of the architecture can be swapped out depending on your requirements. Cheers, Ross MuleSource