Tuesday, March 30, 2010

"Web services != SOA" @ NZOUG's smelliest conference yet

Before your intrigue into the title of this post gets the better of you, I am pleased to announce that we have a guest blogger, Chris Muir, writing for The Red Room.

Chris is an Oracle ACE Director. To become an Oracle ACE, Chris has qualified according the following criteria:
  • Track record of major contributions to the community, technical and/or community-oriented

  • Good communication skills

  • Ability to commit to participation in an honorary (non compensated) capacity
To all these points, Chris' Blog - One Size Doesn't fit all , is one of the most widely-read blogs on JDeveloper and ADF in the international Oracle Technology Network Community and he is one of the foremost Oracle bloggers in Australia. In fact in 2009 he was voted ACE Director of the Year . He is also a prolific contributor to the AUSOUG community.

Chris recently presented to a series of User Group Meetings in New Zealand, Canberra and South Australia and kindly agreed to summarise his presentations in the following guest Blog for The Red Room.

"Web services != SOA" @ NZOUG's smelliest conference yet

Oracle's RedRoom invited me to write a guest post on their blog about my recent travels and presentations. As an Oracle ACE Director I have the chance to attend and present at conferences and events around the world. I specialize in the Fusion Middleware space, with a strong interest and background in the database.

NZOUG's 2010 conference
This March I was privileged to attend and present on Web Services to the New Zealand's Oracle User Group near-yearly conference, as well as present to both the ACTOUG and AUSOUG-SA members on the way home to Perth.

How would I sum up the NZOUG conference?:


Not the usual assessment of an Oracle User Group conference, but this conference on Lake Rotorua, a still active caldera, did take some getting used to thanks to the rotten eggs smell that hands over the small surrounding town.

Yet, this year's NZOUG conference, besides its smell, again proved the NZ event is one well worth attending, for the Oracle content and the friendly delegates and committee members. Of note to anyone interested in presenting or sponsoring the NZOUG event, NZ equalled the number of delegates and sponsors at the reportedly much larger 2009 Australian Oracle User Group conference in Melbourne. Being an ex-AUSOUG committee member I'm well impressed that a country with a fifth of the Australian population punched so much above its weight.

Web service extravaganza
SOAP web services and its newer complimentary cousin REST based web services have a number of significant advantages that organisations and their systems can make use of:
  • The ability to share data in a near real time fashion, with the resulting business benefits of information "now", not tomorrow, next week, or next year

  • To specify clear interfaces and data payloads between systems that can be validated at both design and runtime, to avoid sending or receiving invalid, unstructured data messes

  • Utilise existing, prevalent Internet infrastructure, requiring no dedicated peer-to-peer network infrastructure, avoiding costly network setups

  • Ignoring the other system's underlying technology implementation, who cares if one system is an Oracle system, the other .Net, Java, SAP or similar
Yet given these exciting benefits, the challenge of learning and applying web services can be a hard one for organisations and developers not familiar with its implementation.
At the NZOUG conference, and ACTOUG and AUSOUG-SA events, in order to address this challenge, I had the chance to give delegates a web service introduction, with two entirely different approaches.

The first presentation "Back to Basics: Simple database web services without the SOA headache", a somewhat glib title was designed to draw the predominately SQL and PLSQL proficient delegates at the conference to learn about web services. It's always easier to learn something if you already have skills required to solve a problem. The presentation harnessed the attendees' skills and showed them through the Oracle database how they consume and publish their own web services using the Oracle database's own features.

My second presentation, "JDeveloper 11g Web Services – as easy as 1-2-3" took the more traditional approach to web service development, introducing the contemporary JEE solution using Oracle's own JDeveloper IDE. For some JDeveloper has been a dirty word in the Oracle development circles, with extremely complicated development resulting in steep learning curve and failed system development. But since 10.1.3, and especially the 11g version, JDeveloper like much of the Java arena has tidied up its act and made contemporary Java development easier, a product Oracle is very excited about. My second presentation looked at the simple acts of creating a web service in the IDE through the generation of an XML Schema using a diagrammer, a web service WSDL also using a diagrammer, then finally the Java code via wizards & editors. The funny thing about the presentation is it takes me more time to explain than it actually does to implement the web service – surely a statement that should sell web services itself.

But does SOA have a place?
The nice thing about both these presentations is once delegates have obtained a knowledge of SOAP- based web services, be it through the database's features or JDeveloper, it then allows us to use that new found knowledge as a stepping stone to learn and understand SOA's approach to the broader issues of system integration and a service-based focus. Without the knowledge around the pros and cons of web service development, the leap to appreciating the challenges SOA attempts to address is a hard one. Instead if we provide a learning path between their traditional skills and the new technology, this will address a large segment of people who are essential in getting on board the SOA approach.

The question, will SOA work for you and your organisation is really up to you to decide. Yet, I think it a valid approach to learn, experiment and try out web services first, in order to see if SOA maybe a larger solution that may work for you. And even if it doesn't, web services alone may provide value to your organisation beyond your current implementations.

With thanks to Oracle's Red Room to inviting me to post this guest post, and also with thanks to the NZOUG, ACTOUG and AUSOUG-SA committees for inviting me to present, and the members for attending.

Chris is an Oracle Consultant and Technical Lead and Trainer based in Perth with SAGE Computing Services .

You can follow Chris on Twitter here: http://twitter.com/chriscmuir

Thursday, March 25, 2010

Consolidating Oracle Databases on Private Grids

I've enjoyed the Oracle Red Room blog for some time now so I'm pleased to be finally making a contribution of my own. I'm Roland Slee and after 17 years in sales at Oracle I've recently become the senior representative for Oracle's Database Development organisation in Asia Pacific & Japan. You can learn more about what I do by listening to the podcast below.

I've spent the past two days at the Gartner Infrastructure, Operations and Data Centre Conference in Sydney. The conference themes were "Cloudy, Green and Virtualised". It was an excellent event and I thoroughly enjoyed my discussions there with Gartner analysts including John Roberts, Phil Sargeant and Errol Rasit. I gave a presentation on consolidating Oracle Databases onto shared grids and participated in a vendor panel with Phil Davis from Dell and Gordon Makryllos from APC.

Much of the talk at the conference was focused on virtualisation and cloud computing, topics that remain somewhat theoretical for many customers, so I was keen to offer some very practical ways that customers can adopt an agile, virtualised infrastructure to run their Oracle Databases.

I described in my presentation the way an Australian customer has consolidated more than 300 Oracle Databases onto three database grids, significantly lifting the quality of service for their Oracle systems while also improving DBA productivity and saving money. Oracle Database workloads are found in nearly every enterprise data centre. The Oracle Database is a workload that lends itself to consolidation and virtualisation because Oracle Real Application Clusters allows Oracle Databases workloads to be migrated from scale-up environments to scale-out ones, usually without change to business applications.

The Sun Oracle Database Machine makes this style of consolidation even more attractive because it offers a pre-built, pre-optimised "private cloud" configuration that is ideal for database consolidation as it offers extraordinary performance, availability and agility through its ability to leverage industry standard servers and storage in a transparent, scale-out architecture.

Here's the presentation:

Here's a podcast that summarises the key messages.


So if you're running multiple Oracle Databases on dedicated, scale-up infrastructure then there are great savings to be had by consolidating your database workloads onto a low-cost, agile grid. The Sun Oracle Database Machine is an ideal platform for such a consolidation.

Monday, March 1, 2010

Getting down to business with SOA

This week, I stumbled onto an article in CIO, written by Randy Heffner, VP at Forrester Research. In his article, Randy covers two points that really piqued my interest. The first was with regards the adoption rate of SOA, and the second point was aimed at what CIO’s really need to know about SOA.

While the hype around technologies such as Cloud computing is at a high according to Gartner, SOA has been quietly maturing, resulting in a sound adoption rate (68% adoption by end 2010 according to Forrester), and delivering sufficient business value to warrant 52% of organisations using SOA to expand its use. While my views are not empirical like Forrester’s, the conversations I am having with organisations has shifted from “what is SOA?” and “is it voodoo?” discussions; to “what are the actual steps we need to follow to facilitate the IT transformation we need to go through” and “how do we govern our architecture to ensure maximum service reusability”. I believe that a contributing factor behind this shift in discussion, and an increase in maturity, is that many COTS applications today are service enabled. These pools of available services have allowed IT practitioners to change the way they do application integration, and now they are at the point of asking how to expand and control their SOA initiative. As this is traditionally an IT driven initiative, the drive is towards new development approaches and potentially service reuse. I agree with Randy in that service reuse is only part of the picture and the true power of adopting a SOA approach is not being realised.

So what is the other half of the coin that is missing? I would call it strategic business thinking. Randy eloquently stated that CIO’s need to understand that SOA is about having the capability of designing software (business services) around the business capabilities required by their business. These business services utilise the value of existing applications yet hide the complexity of the underlying architectures. The end result is a business that has the capability of rapidly assembling business capabilities to address their transforming business. Hence the mantra of SOA bringing agility and flexibility to a business. If I played “bingo” on the word “business” in this paragraph I would be a rich person! But that is what it is about. CIO’s need to understand the business strategy, the business transformation strategy and the business capabilities that will be required in the future. By combining a business and reuse approach, the value of a business service will be determined not only by its technical level of reuse, but the value it delivers to the business capability requirements.

So, we IT practitioners should not be asking ourselves the question “how do we make SOA successful?” SOA will deliver value to your business when it is utilised as an architectural approach to develop business capability. If that message gets diluted in any way, you will end up with a costly, moving part in your architecture that does not live up to its full potential.