The Grain of Salt to Take with Oracle Fusion Middleware

 
 
By Lisa Vaas  |  Posted 2005-10-02
 
 
 

The Grain of Salt to Take with Oracle Fusion Middleware


BOSTON—Isnt it nice that Oracle has stopped trying to be the center of the universe, with that whole one common data model, everything-feeds-into-Oracle stuff they preached for the past five or six years?

Theyre beginning to sound just like IBM, actually, with the whole notion of leaving data where it lives, in whatever applications or LDAP systems or databases in which said data has been accustomed to living, and using whatever tools to manage it that customers want.

Leave data where it is. Dont rip and replace the huge investments of time, money and infrastructure youve made. What a concept!

What a change from the past, with Oracle trying to convince everybody that they needed to move to one common data model, with one workflow engine, with one directory, and with applications built in an integrated way around that common data model.

Of course, over that span of time, Oracle was working to integrate 190 general ledgers, e-mail systems, etc., all from acquisitions of distributors and anybody else who looked tasty. Thats a heap of data models to have under your roof, none of them consistent.

So, said President Charles Phillips at the Fusion Middleware road show in Boston on Thursday, that got Oracle brains ticking.

"It was a mess. That was the thing that got us thinking, If were like that, probably a lot of customers are like that," Phillips said.

Hence, the center-of-the-universe approach, and it lasted a good, long while.

The problem was— hello, common sense here?—not everybody could afford to just stop everything cold and migrate to one data model.

The secret to tying all that disparate stuff together, Oracle has found, is not to shove everything into one common data model. No, the secret is actually in the middleware. IBM already knew that. But so did most of us.

So Oracles key strategy now is to reach into systems and control the data, using this whole services-based approach.

After all, just because Oracle isnt the center of the universe anymore doesnt mean it doesnt want to have little strings to yank on every part of the universe.

Anyway, why should Oracle need to be at the center of the universe? At this point, its purchased the entire universe.

Click here to read more about the Oracle buying spree from columnist John Pallatto.

What exactly is different now, though? Phillips said that for once, everybody seems to be marching in the same direction and using the same standards.

"Given thats the premise, we assume ISVs are marching to that same place, accepting those standards, and we assume customers, when they build applications, will use those same" standards, he said.

So the question is, what architecture is needed to do that? Oracles answer is Oracle Fusion Architecture, and its geared to span all applications, not just Oracle applications. "Before, all we could touch was the data," Phillips said. "What about everything else?"

Oracle has defined the key attributes of this architecture as being that its got to be model-driven, its got to be service- and event-enabled, it must be standards-based, its got to be information-centric, and it must be grid-ready.

At the base level is grid, giving security, clustering and information management and allowing scaling with lots of small processors: lots of small Blade servers, say.

Then you get Fusion Service Bus, the main highway for messaging routing, where consumers call services.

A new, improved service bus is expected out shortly, in 10g R3.

Tom Kurian, senior vice president of Oracle Server Technologies, told me in a Q&A at the roadshow that we can look forward to support for new emerging standards, as well as lighter weight, pluggable transport.

Also, well see more capabilities around business rules and data transformation, including the ability to declare a set of rules that make it a lot easier to do routing and data quality services.

Finally, while every enterprise service bus supports the asynchronous model, some use cases require the synchronous model. In the next Oracle ESB, well get a bus that gives us either model.

Next up is the business process orchestration layer, where the BPEL (Business Process Execution Language) engine lives. Thats where all the magic is. Oracle says its got the best BPEL engine on the market, and at least some people think its got that right, given that it bought the right stuff when it acquired Collaxa last year.

On top of that is the real-time analytics, activity monitoring and BI layer, and on top of all that is the unified portal, which can span multiple instances, as opposed to the competitions portal.

Thats the plan. So whats believable, and where do we need to apply the grain of salt?

Next Page: Oracle has a lot of work to do.

Oracle Has a Lot


of Work to Do">

Judith Hurwitz, of Hurwitz & Associates, says Oracles got the portal thing right.

She told me as well that SAP, Oracle competitor par excellence, has a lot of work to do with master data management.

As it now stands, customers have 10, 15, 20 SAP instances—especially those using the older R3 versions.

"You try to manage data across 10 instances of SAP—its really hard," she said.

SAP is trying to get customers to MySAP as a core, running modules on top, so they have one instance working with different applications.

One key problem: People get bogged down in day-to-day management.

Take one SAP customer whos a Hurwitz & Associates client, from the Venezuelan oil industry.

The company used to have 250 people working on its SAP team. Thanks to cutbacks and consolidations, thats down to 50.

"Theyre just trying to keep their heads above water," Hurwitz told me. "Its hard to say, OK, Ill move from my old implementation and embrace this new world."

Thats not just an SAP problem, of course.

Oracles brave new Fusion Middleware world is facing the same issues. Resources are tight and major moves unlikely, particularly given that youre realistically looking at something like a 10-year journey to get to the SOA promised land.

If everyone had a blank slate, it would be one thing. But everybody has existing systems and customers, and nobodys IT organization is getting any bigger, Hurwitz noted.

And as it is, people dont even know what this SOA stuff is.

Hurwitz & Associates did a study last year, of companies that werent small, in which people said "Yeah, were going to services. Were moving to client/server."

They dont understand this stuff, and theres no vendor doing a good job of telling them what this service stuff is.

Because really, you cant just leave data where it is and slap SOA on it like a coat of paint.

You cant just go in and put XML wrappers around code or add service interfaces. All that needs to be done, but you cant start there. If you do, you might as well just give up.

What really needs to be done is analyze the hell out of your business. You have to disassemble it into every moving part.

You must know what all the component pieces are that make up your business. If you dont do that, Hurwitz said, youre running in place.

So were right back to where we were when Oracle wanted everybody to move to one common data model, but nobody wanted to go through the system and business upheaval.

Thats why SAP has a better message, actually, in this case.

Their message is that you have to start with asking yourself what are your business components, and what steps do you have to take to get to SOA.

So. Its déjà vu all over again, but at least Oracle has acknowledged there are other systems in the world, and they arent going anywhere anytime soon.

Lisa Vaas is Ziff Davis Internets news editor in charge of operations. She is also the editor of eWEEK.coms Database and Business Intelligence topic center. She has been with eWEEK and eWEEK.com since 1995, most recently covering enterprise applications; database technology; and RSS, syndication and blogging technologies. She can be reached at lisa_vaas@ziffdavis.com.

Check out eWEEK.coms for the latest database news, reviews and analysis.

Rocket Fuel