In recent blog post, I shared that I had come to Oracle based on the strength of their cloud strategy.
A few snarky individuals tweeted "what strategy?". Haha, the interwebs would certainly be less entertaining without a bit of trolling.
To be fair, I realized I hadn't made my case for Oracle's approach to cloud.
So I decided to do just that.
This is a long post -- longer than most. It is not corporate marketing, it is my personal perspective. Disclaimer: these are (as always) my own opinions, and certainly not vetted by my employer.
Viewer discretion advised: I'm going to be direct as usual, and some sensibilities will be inevitably offended in the process. Apologies in advance.
Why Do I Care?
I have spent my entire working life -- all 40+ years of it -- involved with enterprise IT, mostly on the vendor side.
I am not exaggerating: at 14, I had a part-time job coding BASIC on a timeshare mainframe. Teletype, punch tape and a 300 baud modem.
That sort of thing tended to happen in Silicon Valley where I grew up. Cheap labor :)
As a result, I have grown a deep empathy for those that choose to make enterprise IT their career.
There are certainly easier ways to make a living. I see IT leaders and thinkers continually struggle with competing demands from the business against a rapidly shifting technology landscape. They rarely get the love from the executive team. It's a tough gig.
When I see something that can potentially offer IT professionals a win, I go out of my way to say "look at this, it could be big for you!". I can list off all the different technologies and strategies I've been smitten by.
Most of them became real things -- at least, for a while.
I was smitten early on by the promise of cloud, but became disillusioned over time. Many current forms of cloud have not proven to be that big win I had hoped for, especially for enterprise IT.
Most public clouds are inherently difficult for enterprise IT to consume. They don't work the way enterprise IT needs them to work. So adoption is limited; and it doesn't change the game enough to be meaningful. SaaS would be a clear exception here.
I think of most private cloud as a head fake -- "cloud without cloud". It's usually infrastructure only, it's still purchased and operated as plumbing, and there are few compatible public cloud offerings.
It doesn't change the game enough to be meaningful.
As I looked across the IT industry, I was initially quite despondent. Sure, various vendors were offering piece parts that could be potentially cobbled together, but that isn't the greatest answer. Cloud should be something you use and build on, not hand-craft and support forever.
I like driving cars. I don't enjoy selecting automotive components, attempting to assemble and integrate them, and supporting them when they break. Especially since all I mostly do these days is drive to the airport.
The business is starting to look at IT the same way. Why can't I just consume what I want to consume? Who is this army of well-intentioned IT people attempting to build and support stuff, and continually struggling in the process? Why don't we just move to a better model?
And -- unfortunately -- those business people are right.
The Oracle Approach
At a high level, here is what I find so compelling about Oracle's approach to cloud.
First, there's a complete -- and integrated -- public cloud offering that works the way enterprise IT works.
There's integrated and extensible SaaS.
There's a PaaS platform aimed at enterprise developers with a dizzying set of capabilities.
And there's a broad range of IaaS services that covers the enterprise IT spectrum.
You're free to compare subsets of the public Oracle Cloud with other alternatives, but there is no direct comparison for the entire offering.
Second, there is a full line of on-prem systems engineered for database, applications and analytics. They do things other systems can't do.
And they come with precise equivalents in the public Oracle Cloud should you need a different consumption option down the road, which I call "cloud insurance".
Cloud insurance? Yeah, you don't want to be "that guy" when new IT leadership decides that everything is going to some sort of cloud.
Again, you won't find any on-premises infrastructure vendor doing this.
And, finally, Oracle has figured out how to bring subsets of the public Oracle Cloud into the data center, behind your firewall.
Same functionality, same operational model, same pricing model, etc.
Cloud machines: a public cloud model delivered on-premises: 100% compatible with the public Oracle Cloud.
No other vendor has figured out how to do that, either.
Put it all together? A complete and modern full stack optimized for enterprise applications and workloads. One architecture, three cloud consumption models, your choice.
Use it to run aging legacy, or build cloud-native: your choice. Add in workload portability with other clouds, and it starts to look pretty appetizing.
I don't think any vendor can make any one of these statements, let alone all of them. So when I get on my soapbox and start saying "hey, this could be big for you!", there's some reasonable substance behind the claim.
But I think I owe you a deeper explanation of why each component of the Oracle approach is so compelling through my eyes.
A Business Process Platform View Of SaaS
A narrow view of software-as-a-service holds that certain business functions aren't strategic. Rather than invest resources in either customized or on-prem solutions, simply consume an external software service for what you need. Packaged, black-box functionality is all we need.
The broader view is that most organizations thrive on their business process DNA. The better companies are always investing in new processes, and improving existing ones. Moving to a SaaS model should support that strategic view, and not inhibit it.
What sets Oracle SaaS apart from other SaaS offerings in my view are three things: (1) a broad set of integrated core enterprise functionality (2) rich, ready-to use business process IP, and (3) the ability to easily extend and integrate functionality as needed.
Put differently, I tihnk of Oracle SaaS as a business process platform, and not a point solution.
A Developer Productivity View of Platform As A Service
While there always will be a strong desire to run shrink-wrapped software, that's not always enough. Enter the enterprise application developer. Or perhaps a small army of developers?
Developing and customizing application software is an expensive proposition. The focus here is to make these folks as productive as possible in their core task: delivering powerful applications.
Yes, there are all the developer tools you might expect for the coding bits (plus the vast world of open source tools) but there's also a very long list of powerful, ready-to-use services that can be easily integrated into even more powerful applications: mobile, analytics, IoT, etc.
Where I think Oracle stands apart is their focus on the unique needs of enterprise application developers, and giving them the tools they need to do their job far better.
An Enterprise IT View of Infrastructure As A Service
Start talking about public cloud IaaS, and people think "AWS". After all, they pretty much invented the category over ten years ago. And they have been quite successful, no argument there.
That being said, it's fair to say that AWS is a poor fit with the needs of most enterprise IT functions.
Not all applications can run well in a virtualized x86 environment, for one thing.
If you've invested heavily in people and process around control planes (security, monitoring, etc.) none of that works well with AWS.
Not every app can go to a public cloud, there's still a strong need to run things on-premises, ideally with the same architecture.
For dedicated and predictable workloads, AWS can be awfully expensive. And if you're not paying attention to consumption, "self-service" can quickly become "all you can eat" and blowing out the budget in the process.
I argue that there's clearly room in the market for public cloud IaaS that works the way enterprise IT needs it to work.
And that is what Oracle is building.
Choose from a wide variety of compute models: elastic, dedicated, bare metal or our engineered systems. Run Linux, Windows or UNIX. Run legacy, run virtualized, run containers. Get workload portability at the database, VM or container level.
Larry Ellison made it very clear during his Oracle Open World keynote that he intends to directly compete with AWS on both price *and* performance. And I see clear evidence that Oracle has more than enough firepower to deliver on that promise.
Now, On To The Data Center
Despite all this cloud talk, today's reality is that the vast majority of IT spend still happens in the data center. And that's not going to change next week.
Oracle's infrastructure is all about enterprise applications: databases, application logic and middleware, and analytics. We use the term "engineered systems" to describe our full-stack integration of hardware and software.
For people who want to build customized on-premises environments, we'll sell them the same components we use in our public cloud: operating system and hypervisor, servers, storage, data center fabric, etc.
Caution: some assembly is required.
For folks who want a simplified out-of-the-box experience, we offer purpose-built appliances: one each for databases, application logic and middleware, and analytics.
For folks who need something a bit more extreme, that's where products like Exadata, Exalytics, Exalogic and SuperCluster fit in. For what they do, no one does it better. That's what happens when you can engineer hardware and software to work together.
So, what's the tie in with public cloud?
Our on-premises technology has precise equivalents in the public Oracle Cloud. I think of it as "cloud insurance".
Sure, you want an on-premises solution today, but what about down the road?
Isn't it nice to have a precisely equivalent public cloud option -- just in case?
Bringing The Public Cloud To The Data Center
There are so many situations where a public cloud model might look attractive, but can't be considered for any number of reasons: regulations, latency, etc.
If you can't come to our cloud, we'll bring our cloud to you.
That's the idea behind cloud machines: subsets of the public Oracle Cloud delivered as a service in your data center, behind your firewall. It's managed and monitored by the same people (and using the same tools) as our public Oracle Cloud.
No one else in the industry is doing this today. I do read that Microsoft is working towards something similar in the future. Hard for me to compare with something that doesn't exist yet.
The current Oracle Cloud Machine is integrated PaaS/IaaS targeted at application landscapes. The PaaS/IaaS services are identical to those delivered in the public Oracle Cloud. Many aspects of the pricing are the same as well.
At Oracle Open World, we announced that we'd be delivering an Exadata Cloud Machine, aimed at providing the same database-as-a-service capabilities as we offer with Exadata in the Oracle Cloud. We also stated our intent to do the same with open analytics, with a Big Data Cloud Machine.
To be very clear, a cloud machine is not a private cloud. It is a public cloud model, delivered in your data center, behind your firewall. You use it; you don't buy it/build it/maintain it. Just like a public cloud -- although certain minimums apply.
Why I'm Such A Big Fan Of The Oracle Approach
One of the reasons that I think there's been so little adoption of public clouds by mainstream IT organizations is that there's so much that has to change all at once. In most situations, it's an all-or-nothing proposition.
Burn the boats, we're going to the cloud!
In most cases, your applications have to be converted and tested to run in a given public cloud -- or you have to move to SaaS. Any operational procedures or tooling you have invested in has to be re-invented.
And, once you're there, you're not coming home -- even if you need to.
Put differently, if you could somehow break the cloud proposition into manageable, easier-to-consume pieces, more IT shops could make progress towards that goal.
As an example, imagine a public cloud that didn't require enterprise applications to be converted or rewritten. One that supported legacy monolithic applications, current multi-tiered applications as well as the new generation of container and microservices apps.
As another example, imagine that you'd stand a decent shot of simply extending your chosen control planes and policies into the public cloud.
As a final example, imagine that you could easily consume either on-premises or public cloud using the exact same architecture. And you could have straightforward workload portability to other clouds and other architectures.
That would be an entirely different proposition, now wouldn't it? Far easier to consume than burning the boats and starting all over again in a new world.
Move your applications without rewriting or conversion. Extend your control planes of choice into the public cloud. Consume a public cloud model in the convenience of your own data center. Move workloads between the data center and your choice of cloud. Rich PaaS and SaaS functionality that builds a platform for business innovation.
That's why I came to Oracle. I believe that cloud will eventually change how enterprise IT will be done. And Oracle was the only company I could find who had all the pieces -- and the commitment -- to make it happen.
Everyone other vendor I looked at either had a partial answer, or often no answer. Unless you'd like to assemble something yourself? Hint: it's harder than it looks.
Now that I'm over a year into this, I am a stronger believer than ever. The story and the execution keeps getting better and better. When I put this story in front of an enterprise IT group, they immediately realize it creates entirely new options for them.
Integrated SaaS/PaaS/IaaS delivered in the public cloud.
Engineered systems optimized for database, application logic and middleware, and analytics -- with precise equivalents in a public cloud.
And the tantalizing notion of a cloud machine: a public cloud model delivered in your datacenter, behind your firewall, as a managed cloud service.
All aimed at enterprise IT organizations, and demanding enterprise application workloads. You know, the hard stuff.
So far, so good.
But we're in early days. The big shift from traditional on-premises models to cloud models (whether delivered in a public cloud or in the data center) has just begun. Most of the market is still up for grabs.
I remember a time when we all thought that AOL would rule the internet, and we'd all be using Netscape as a browser. Things change.
I think I've chosen an excellent spot to watch the action.
---------------------------------
Like this post? Why not subscribe via email?
Agree. 90% of our applications don't need elasticity, worldwide redundancy and availability, etc. They need pro-active monitoring from metal to GUI and someone to yell at when its slow. Aside from Saas (you can yell at them), very difficult to do in a public cloud only solution...except for small shops perhaps.
dreez
Posted by: Michael Endrizzi | November 12, 2016 at 02:12 AM
"Cloud insurance? Yeah, you don't want to be "that guy" when new IT leadership decides that everything is going to some sort of cloud.
Again, you won't find any on-premises infrastructure vendor doing this."
Check out BlueData. Doing it for Big Data workloads. Started with on-premises and now for cloud. You guys should be looking at them closely.
Posted by: Andy | December 09, 2016 at 11:10 AM