Analysis Remember The Big Switch – the book by Nicholas Carr which said that IT would become a utility-like service delivered through a socket in the office wall?
We know it's only patchily working for business users, as public cloud adoption is going up but is still not a serious threat to on-premises giants.
Both camps, the public cloud pushers and the on-premises-first group, can say they are doing OK.
We spoke to Oracle cloud evangelist Chuck Hollis on why his firm thinks you should shift off-premises and whether it's eating its own dog food.
El Reg: Tell us about the rise of the public cloud as you see it.
Chuck Hollis: I too thought the vast majority of on-premises IT vendors would face a bleak future: fighting harder for their share of a shrinking pile of table scraps as they become increasingly irrelevant. None of the on-premises players look to have good prospects: Cisco, Dell, HP et al. Not sure if their customers were aware of what the future would hold, or not.
I thought the only way an on-premises vendor could have a relevant cloud story was to actually have a cloud.
El Reg: Why was this?
Chuck Hollis: When I was at EMC, we thought of all sorts of clever ways to make our products work with other vendors’ clouds, but realized we were just making it easier for our customers to become someone else’s customers.
I think that same thought has come up at NetApp, HP, VMware et al.
My boss Dave Donatelli was also making dire predictions way back when (July 2015): “If you’re in the infrastructure business the cloud has fundamentally changed the industry. Hardware companies are trying to sidestep the issue with marketing and by attempting to position themselves as ‘cloud arms dealers,’ but the fact remains—any hardware vendor without a successful public cloud business will face significant business challenges.”
El Reg: Why did Oracle go into the public cloud business and not the others?
Chuck Hollis: Why aren’t the traditional on-prem vendors getting into the cloud game? The short answer: they can’t.
El Reg: Why not?
Chuck Hollis: I think that’s for two reasons.
First, the shareholders of any publicly held company won’t allow it to spend the required many billions of dollars and depress earnings for the five or more years it takes to build a profitable cloud business.
The antidote? Founder money. Amazon, Microsoft, Google, Oracle et al have founders who own a significant amount of stock. So they can prevent activist investors from storming the citadel.
El Reg: And the second?
Chuck Hollis: You need a natural franchise. Amazon got first mover advantage, congrats to them, they picked vast swaths of developers who were under-served in the market by the on-prem crowd. They build a franchise from scratch, good for them. Microsoft has their natural desktop/workgroup/collaboration franchise. Oracle has enterprise database and the applications that use them.
Note that VMware could be argued to have a natural franchise (second point), but got tripped up by the first (it’s way expensive). HP? Dell-EMC? Cisco? Even IBM is hamstrung by these two points, despite a continual struggle.
... A lot of what passes for “strategy” these days from the on-premises vendors I see as simply buying time.
El Reg: What about mission-critical applications not moving to the cloud?
Chuck Hollis: [There's] a critical mass of enterprise applications that form the gravity well for enterprise IT.
You know, the important stuff. In the enterprise IT world, those enterprise apps sit at the very centre of the universe. It’s hard to envision any important business process that doesn’t touch enterprise apps.
The real game in enterprise IT are those enterprise apps, what’s the play?
El Reg: What - and how - is Microsoft doing here?
Chuck Hollis: I think Microsoft did a solid job with Azure. If anything, I personally see it more of a relevant competitive threat to Oracle’s ambitions than anything AWS or Google could muster up. Too soon for me to have an opinion on Huawei.
Microsoft has led their on-premises franchise right into their cloud and now they’re trying to expand. But I think they’re missing something important: [the] critical mass of enterprise applications. ...
Strategically, I find this interesting, as most everything important in an enterprise IT landscape revolves around those beefy, mission-critical apps. All data captured leads there: web data, IoT. All business decisions are made there: analytics, big data, ML, etc. All actions within a business are driven by enterprise apps.
El Reg: And Amazon?
Chuck Hollis: The mainstream press points at Amazon as the “market leader”. True, by revenue, mostly IaaS. But there’s a whole lot of enterprise workloads left to go to the cloud. And if you’d like to get an IT manager riled up over dinner, ask them how their AWS experience is going.
Bottom line: Amazon built a successful cloud, but it’s not designed to work the way most enterprise IT organizations need to work. That was never the intent. As a result, a lot of time is spent piecing things together and making them work, sort of as we did in the on-premises era.
It’s certainly a candidate if you’re building a greenfield application, but that’s not most of enterprise IT.
El Reg: OK, you're going to tell us Oracle is doing it better, so tell us.
Chuck Hollis: Oracle not only has its enterprise application business, it also has been the preferred database for so many ISV enterprise apps...
Oracle’s view is simple: SaaS suites win in the long term.
SaaS starts with ERP (after all, financials are the core of any business), and then expands to other areas of the business. Right now, Oracle is doing quite well in SaaS ERP, which is a good thing if you look at the broader context. We’re also starting to see our successful SaaS ERP customers starting to add human resources, customer experience and so on. So the strategy appears to be working.
We have this belief that things that are designed to work together will ultimately be better than lashing up a bunch of separate SaaS providers, as people are discovering...
If people don’t want to move to SaaS, we lift and shift their enterprise app landscape to our cloud.
El Reg: And if business really doesn't want to move enterprise apps to the cloud?
Chuck Hollis: For folks who want a cloud experience but don’t want to move to a public cloud for any number of reasons, Oracle offers Cloud at Customer, which is essentially our public cloud delivered as a service in the customers’ data centre.
It’s a bit different than Azure Stack in that Oracle provides the complete solution: hardware, software, operations, etc...
El Reg: Has Oracle moved its enterprise apps to the cloud?
Chuck Hollis: [Oracle] has re-implemented its entire business on a modern cloud platform – SaaS, PaaS and IaaS. Remember we’re talking a ~$200bn market cap company here – no easy trick. The fun thing is that I’m part of a project to document the before and after around a whole raft of internal business metrics. The comparison is stunning, to say the least.
Chuck Hollis's point of view is self-serving but also logical.
We have two massive pure play and general public cloud IT service companies, in the form of Amazon and Google, and one hybrid one – Microsoft with its Azure service paired with its on-premises software – all hoping you'll trust the public cloud with matters business-critical.
Then there are the legacy on-premises IT cloud naysayers, who want you to keep mission-critical computing on-premises and use the public cloud for less important and/or temporary work. The opposing on-premises IT-first camp includes Cisco, Dell, HPE, NetApp, Nutanix, Pure Storage and Vantara plus a whole host of other players. Enterprise won't move all in to the public cloud, they say, because it's insecure, unreliable, costs too much in reality and is complex.
The legacy on-premises crowd claims to make your on-premises IT experience cloud-like but without the disadvantages. They say they can offer cloud consumption models, and software-driven, flexible infrastructure control for your on-premises kit – a me-too pitch.
Neither the public cloud pushers nor the on-premises group have arguments that have won the day yet. Neither has delivered a knock-out blow to the other.
The arrival of such a blow will be indicated by public cloud business use starting to flatline and then maybe decline, or its opposite. ®