Ellison: the Internet is the last big shift

According to Larry Ellison, the move from client/server to the Internet is much bigger than the shift from mainframes to client/server ever was.

“Every company always says it’s an Internet company. They are born-again Web companies. But we’ve been changing our software for four years – it’s taken us a long time to figure out how to do this,” Oracle Corp.’s CEO and chairman said in a keynote speech, which was patched in from the company’s headquarters to attendees at its recent OAUG conference in Orlando, Fla.

Ellison’s video image was beamed larger than life from Redwood Shores, Calif., onto huge TV screens where some 8,000 attendees at the conference were expecting to see him appear live.

Ellison said Oracle has been going through a very difficult period of technology change – by abandoning client/server and moving to the Internet. “If a company misses one of these shifts, it disappears. So we are very fortunate that we caught this early,” he said.

“I don’t believe there is anything after the Internet. It’s the last big paradigm shift.”

Up until now, Oracle hasn’t done a very good job of turning transactional information into useful data necessary for running a company, Ellison admitted. “Though there are a lot of reasons why that was the case, perhaps the single worse reason was the fact that we had fragmented all of our information,” he said.

He used the example that Oracle itself has ended up with 70 separate HR databases spread across its global sites.

“That means we had to buy 70 computers, hire people to run all those computers. And here I am, CEO of the number one company in the world for managing information and I want to ask a very basic question from our fabulous HR system: ‘How many people work at Oracle?’ And I don’t know,” Ellison said.

“In order to find this out, I have to go into 70 different databases in 70 separate countries and add all this up. And an even more important question is ‘How much did we sell today?’ Now, executives really like to have this information. But we’ve so dreadfully fragmented our information that we don’t know what’s going on in our own company.”

Ellison pointed out that distributed computing really ends up as a model for distributed complexity.

“It wasn’t very long ago when Microsoft was talking about little NT servers everywhere, little databases everywhere. This is a really bad idea,” he said as the audience laughed.

“You can’t slice and dice your information. You can’t divide it up by country, and you certainly can’t divide it up by office.”

Gathering all your databases into one or two data centres helps a little, he said, “but until you consolidate the databases themselves, until you put all your HR information in one database, until you bring all your order entry information into one database, all your customer service information into one database, you really can’t understand your fundamental business. You’ve got to know what the status of your customer is.”

Being an e-business is all about sharing information, Ellison said, and automating your supply chain.

“It’s great to talk about doing that. But until we build a single, unified system for our company, until we consolidate that, it’s kind of silly to talk about building an automated supply chain.”

Almost no one is using parallel server technology with Oracle applications today, Ellison said, but he predicted that within about a year “most people will be using parallel server.

“Most 11i applications will run parallel server, and split the load between transaction processing and query processing. [This way,] the query machine doesn’t interfere with the transaction load at all. We think that will give you tremendous additional scalablity.”

During a question period following the keynote, one customer asked Ellison why Oracle Business Applications Release 11i had come out later than originally planned.

“We decided to delay 11i for a couple of reasons,” Ellison responded. “We decided to put more capability into it, and also we knew there wouldn’t be a lot of people implementing 11i around the Y2K deadline,” he said.

“So as we got closer and closer to releasing something around Y2K, we thought it made sense to delay a little bit longer. We started to sneak in a few more features – we felt that was the right trade-off to make.”

Ellison also admitted that there were a few “unpredicted technology slips” that served to delay the product’s rollout.