Oracle today officially launched its second Australian data centre, tapping into familiar concerns about data sovereignty and performance. But why are its two data centres running entirely different platforms, and can Oracle overcome its familiar history of messy integration when it comes to applications?
Data centre picture from Shutterstock
Oracle president Mark Hurd (CEO of HP in a more scandal-ridden period of his career) officially launched the new centre at a press event in Sydney this morning. Oracle already has one data centre operated by Harbour MSP at Darling Harbour, which is largely used for Oracle's managed CRM offering; the second facility is run by Equinix and located in Alexandria in Sydney.
Hurd pitched the Australian centre using two familiar cloud memes: the demand from many customers to have data hosted onshore and a growing demand for cloud options. "In many cases as you build out the cloud there are issues around data sovereignty so we think it's important to take care of that issue, which we are," he said. "There's always a concern about 'where is my data?' That can be true if you're an individual, a company or a country."
As we've noted before, the question of whether particular types of company are actually required to host data onshore (performance considerations aside) is far from black and white. Local operators typically suggest that it is, but there's a clear element of self-interest. (Oracle itself has several local customers happily using cloud services from its US data centres, although that wasn't a point it gave much emphasis to at the launch.)
The second argument for local data centres is raw demand. "The incremental capacity of the second data centre is for increased demand and because we are bringing on new capacity," Hurd said.
The reality appears a little more complicated, and the new capacity is arguably the most important feature. While the Harbour MSP site is offering hosted CRM, the new centre offers Oracle's range of Fusion applications, which includes several other familiar enterprise verticals (HR, ERP and sales automation systems). It also offers Oracle's platform-as-a-service (PAAS) products, which like Fusion didn't exist when the first centre was established. Effectively, the two centres are competing by offering different products, rather than acting as extensions of each other,
If that doesn't sound very cloud-like, Hurd is unapologetic. Oracle is just as happy to sell on-premises equipment, he said. It's a little difficult to escape the impression that cloud is being offered because all the rivals are embracing it, rather than it being a core technology.
Given that Oracle's new centre is offering a range of packages, one obvious question arises: just how well will they actually work together? Oracle has long supplemented its applications portfolio through acquisitions (RightNow being one prominent example in the cloud space), and Hurd describes acquisitions as "a form of R&D". However, Oracle hasn't always done a good job of actually integrating its acquisitions into a single platform; its acquisition of PeopleSoft and Siebel are obvious examples. So I asked Hurd: what is Oracle doing to avoid that kind of debacle with Fusion?
"Any time you have acquisitions by definition the company you buy won't use the exact same technologies and development environments," he said. "There's always a period of maturation. I think that we're doing a pretty good job."
That said, Hurd argued that Fusion would have fewer problems than previous integration attempts. :In the old days, some of our old apps were built on their own pieces of middleware. All the new applications are built on the common Fusion middleware. That gives us two significant technical advantages: integration and extensibility."
One issue Hurd wouldn't touch on was pricing: local data centres often have higher pricing than their US counterparts. Asked by one journalist if that would continue, Hurd dodged the issue. "I won't make any commitments around pricing because you'd probably write it down."