It's a few years since I was last at Oracle OpenWorld so there was a genuine sense of catching up on my part. But then there was a strong sense of familiarity in the 'we're number one at everything we touch' drumbeat that pervaded many sessions.
As I said in an earlier story, this gets a bit wearing after a while, even when it's tinged with a soupcon of mischief. For example the slide claiming 19 out of 20 of the top SaaS vendors run on Oracle left some of us wondering which one doesn't. It's the kind of detail that doesn't matter to many, but which matters a great deal to the analyst peanut gallery who enjoy nothing better than a geeky drinking game. Or the claims that none of the other SaaS vendors has a platform [like Oracle's.] Of course they don't but that doesn't stop feathers being wrinkled. When I jovially mentioned this to Zach Nelson, CEO NetSuite, he said: "I'll have to show him our latest and greatest then." Quite.
Regardless, Oracle's clever and unmistakable messaging says to me: 'If you're running a Salesforce.com, NetSuite et al - you're really being run on our stuff.' The more subtle subtext goes something like this: 'If you're contemplating a shift to [name your aaS here] then we're the folk to call up.' Does it stand up?
Debra Lilley, VP Certus cloud services and a long time sounding board for me on all things Oracle related put it very well when she said:
Direction is one thing, but roadmap is another - and it's not all there on the latter.
She for example would like to see Oracle make firmer strides on PaaS and specifically business process management. Why? Her clear impression is that Oracle is happy to let smaller SIs develop industry specific or market specific functionality that in turn can be productized.
We don't mind building for one customer if we have to but it makes a lot more sense for us to build for a market.
Of clouds and Oracle cloud
Then there's the question of what does 'cloud' mean to Oracle? I asked Steve Miranda, EVP applications development Oracle if he can point me to the infrastructure URL. The one provided offers a wealth of services and sure enough, you can buy pieces of the compute and storage parts in much the same way you can on Amazon.
But as Matt Asay discovered, it's not all quite there because when it comes to buying the database as a general purpose service, Oracle loses the plot somewhat. Asay takes a harsh view, arguing that:
All of which leads developer Jeff Waugh to channel The Princess Bride:
— Jeff Waugh (@jdub) September 30, 2014
It doesn't have to be this way. Just ask Microsoft.
Asay then goes on to discuss how Microsoft is doing a credible job delivering cloud services, with the clear critique that Oracle doesn't. But then Frank Scavo reminds us that Oracle Openworld is first and foremost a sales event. In those terms and in the context of Oracle customers, any expectation of precision on technical matters in the public keynotes is moot. That wont please the cloud purists but then this wasn't their event.
Customers and mixed messages
As I listened to customer panels, partners and executive sessions, it became clear to me that whatever anyone thinks about the actualité of Oracle's cloud credentials, it was largely the right message for the right audience. While many of us in the commenting peanut gallery get frustrated for one technical reason or another, Oracle customers are very much like SAP customers; they are often very large and want to move at their own pace. In Oracle's case, it is clear that security weighs heavily on their minds. How else do you explain Ellison's multiple mentions of the 'S' word?
But then you also have to wonder. On day one, Ellison touted switching Oracle databases from an on-premise to cloud deployment with 'the push of a button.' It turns out not to be quite like that as Ellison demonstrated live on stage during his second keynote. It's more like a short, multi-step process. When asked for my view I opined that it would be hard for a child to *not* get those instructions. Simple? Yes. Delivered. At least at this superficial level.
There was one telling moment when Ellison said:
This is a live demonstration, it could break at any time.
Beta software anyone? Yes. When you trawl through the offerings, it is clear there is still work to be done and no-one is denying that. But I give Oracle kudos for providing a very clear and unambiguous approach to the aaS store.
What else did we learn?
In my last story I wondered whether Oracle is readying for a mid-market play. This was a topic I brought up with Steve Miranda. He told me that while Oracle plays best with large customers, he'd like to see a stronger presence in the sub-$1 billion market.
Separately, Debra Lilley said she believes that Oracle's cloud offerings now mean that tier 2 SIs can work with companies of around 500 employees. That will serve as a warning shot to the likes of Infor, which wants to serve what it sees as an underserved market, and Workday, that routinely picks off those sizes of organization.
Lilley seemed genuinely encouraged by her interactions with Oracle's engineering teams and believes that subject to some caveats around the line Oracle draws in the functional sand, cloud - as in configurable Fusion apps - provides her business with a very good opportunity in a segment that Oracle would likely not go to directly or through its tier 1 SI partners.
Big Data SQL
One interesting and tantalizing snippet came towards the end of Ellison's second keynote that is worth mentioning. See image below. Surprisingly, I heard Ellison acknowledge Hadoop and other types of database as credible and he went as far as talking about co-existence. In Oracle terms, this means ensuring that data can traverse the entire panoply of databases that might be used in webscale operations.
This is interesting because this year I have seen the world bifurcating into two distinct camps: those who believe in the ERP world as the center of the universe and those who see new classes of application as where business critical information lives, albeit in often very lightly defined processes. A good example of the latter is the MyDealerLot story I discussed the other week.
Ellison posed an important question when he asked how you get from one place to another inside the whole database landscape while remaining performant? This is a real problem because modern, non-relational database systems often work way faster than relational systems, even when gobs of RAM are applied to the former. Why does this matter?
If you take the view as I do that sooner or later you need to pass data between systems of record and systems of engagement that use Hadoop, MongoDB, Cassandra and the like then you need to find a way of ensuring that nothing in the overall landscape serves as a bottleneck. It is incredibly expensive to get this done at web scale using in-memory column stores. On the other hand, disk is cheap but slow.
Oracle believes it has found the answer in Big Data SQL.
Right now, that technology is a tad limited and it is not the only game in town. But for Oracle, it also means that those skilled in SQL preserve their value when working between relational systems and Hadoop/noSQL.
Ellison pitched this as 'introducing' although the technology was launched in July. Even so, it's an interesting development that serves to bolster the Oracle/cloud story.
Kinder, softer Oracle?
One question that was left hanging in the air - are we seeing the start of an emerging kinder and softer Oracle? The old joke goes: Oracle hates everyone and especially its customers. But you wouldn't have got that impression from listening to the panel of eight customers Mark Hurd, CEO Oracle brought out on stage. Or the panel of customers Steve Miranda wheeled out. If anything, I'd say that the customers we saw were largely OK with what the company is doing. How do I know?
I spent some quality time with Bob Evans, our main internal sponsor at Oracle. I congratulated him on getting executives 'trained' to not lead customers to predictable answers but let them speak for themselves. It is a topic dear to our hearts at diginomica because we know that customers speak far louder than a gaggle of analysts.
Such things always carry risk because you can never quite predict what the customer will say - something Evans knows only too well from his days as a technology journalist. It worked well as a convincing exercise in communication for which Oracle deserves credit. Now that Oracle has proven the case, I expect to see competitors follow suit. If I am correct, then that is a very good thing.
Don't get me wrong. Oracle is well used to playing hard ball with the best of them and will think nothing of tossing an out of line journalist or analyst into the sin bin until such times as Oracle thinks they've served their time.
From my conversations and the invitations for follow up, I sense that Oracle is learning that in a world of extreme transparency and a thousand voices capable of dissent, those sledgehammer policies no longer work as well as they might have done in the past.
Like their software, Oracle has a ways to go and there is much more detailed due diligence needed on individual topics and technologies. In some cases, mixed messages need clean up but you can see where they are going, understand it and be prepared to wait for the story to flesh out. As Ellison concluded: 'We've just started.'
Endnote: During our discussion, Debra Lilley said she raised the question of influencer relations/communication with Oracle communications. It's a great idea and one that I know has informed SAP in a number of areas. From my perspective, the value of such sessions would come in getting the inside skinny on technologies in advance of events like OpenWorld plus the free flow of ideas and observations from those of us who talk to customers and partners on a near daily basis. This is important because as Lilley also pointed out. She had been in three full days of talks and workshops prior to the event and even then had a full 'dance card' as we term our endless foot shuffling from one session to another.
Disclosure: SAP, Oracle, NetSuite and Workday are all premier partners at time of writing