A serverless, blockchain world? Yes, but let’s get Oracle customers cloud native first…

Oracle cloud native

Chad Arimura has a vision of a world where blockchain and serverless computing combine so that functions can be downloaded from a marketplace and run on any cloud. If that sounds far-sighted, his boss, Larry Ellison stated at this year’s Oracle World conference that the vendor’s second generation cloud is already protected by Star Wars style cyber defence and autonomous robots.

It’s heady stuff, but Arimura, speaking to DevClass ahead of his keynote at Serverless Computing London,  seemed admirably earthbound, talking about how his unit within Oracle is working to develop a Lambda competitor that will make going serverless a lot less scary for companies running the traditional back-end applications that Oracle specialises in.

That competitor is the Fn Project. The technology was originally developed by Arimura’s startup Iron.io, which was sold to private equity a couple of years ago. However, most of the engineers, and Arimura, moved to Redwood Shores, as part of the Oracle Cloud Infrastructure group. His unit is called the Serverless Organisation, and it works both on Fn as an open source product, and as a managed service offering within Oracle’s cloud service

The project received a coming out party at last week’s KubeCon, where Big Red unwrapped its Oracle Cloud Native Framework, which it described as “a cloud native solution that spans public cloud, on premises and hybrid cloud deployments.”

Does the world need another serverless technology? Well, as Rupert Murdoch once said, “a monopoly is a terrible thing unless you have one.”

Arimura obviously thinks so, particularly one that can capitalise on Oracle’s “strategic advantage” in the shape of large enterprise customers, with large application installations and the Java community, with “real seamless elegant integrations with the Serverless architecture.”

“We lead with open source and then we’re creating really great cloud services based on open source projects and we’re starting with function as a service and we’re looking at some other services as well for the future,” he told us.

What do Oracle customers want?

Arimura is realistic about the state of play around serverless today, dominated by Lambda, with Google and Microsoft’s touting their own offerings, and then those under the CNCF banner.

As Arimura explains it, “An enterprise wants to launch its own infrastructure as a service. It wants to launch a platform based on Kubernetes then it wants serverless on top of Kubernetes and that’s where Fn Projects comes in.”

“We also manage Kubernetes (through the Oracle Kubernetes Engine service), we also manage the Fn project as Oracle functions. So you get this nice managed experience on the Oracle cloud and you get the same api and the same developer experience on your own hardware or on another cloud if you want as well.”

While Oracle has undoubted prowess in large enterprises – or depending on your point of view, an iron grip on its legacy customers – there is always debate over whether those organisations are able to compete with smaller, or simply more nimble companies who can quickly exploit the assume advantages of the new architectures.

Arimura intimately understands these issues, saying that a year and a half ago he was running a startup himself and the sort of beefy applications Oracle provided simply weren’t on the agenda.

Nevertheless, he says, “Oracle had the opportunity to help bridge its customer base into the cloud and our enterprise customers are looking to us to shepherd them into managed cloud.”

But simply shifting onto someone else’s tin isn’t the end game, he says, and those “traditional” organisations are also looking longingly at the new tools.

“They want to use Kubernetes, they want to use serverless, they want to use bare metal instances.”

So the aim is “an offering that’s highly optimised for the environment they want to run in, run next to the applications they run on Oracle, and integrated with the applications they use with Oracle.”

The future is autonomous

At the same time, “Developers go from startups to enterprise, from enterprises to startups, they take the tools and the knowledge that they have. You can’t ignore one market for the other.”

“We have some pretty amazing things coming on the market like autonomous data warehousing, autonomous transaction processing,” Arimura says. “And then we have the open source. Use serverless computing but do it in an open source way, container native, and these type of advantages will hopefully give developers a reason to at least check out how much faster they can run things on Oracle, how much cheaper they can run things on Oracle.”

And where will serverless go as Big Red unleashes this new services? “Serverless is square in the target of where the market is going. We’re moving the abstraction layer up. It’s like building on the promise of PaaS. Developers simply don’t want to worry about stuff under the hood.”

“I think we’re past the hype cycle and I think we’re moving onto to ‘ok, how do we use this in our organisation.’”

The lowest hanging fruit for this is things like “Devops type tasks, just simple kind of integration workflows…If you want a server you can run a Lambda function that automatically check whether there are certain tags attached to the server every time an object goes into your storage.You want to do some manipulation of that object you’re going to kick off a function to be able to process that image or video”

But setting up the infrastructure to do this is something that most developers don’t want to deal with. “Those are patterns that you have to really work at to understand. We’re now making this much easier for developers – in fact, even non-developers can launch functions.”

Which is often a cue for concerns about making things too easy for (non-) developers, and whether someone needs to start imposing a little order around here.

Arimura, for his part, is involved in the Cloud Events Serverless Working Group. “I think serverless is all about events, which is data, something happened, and we want to do something to what happened….The first big body of work was cloud events – that is an attempt to standardise event data across the wire.

Did we mention blockchain?

“We basically described some meta data and context of what an event should look like. and then you can put the event payload in there. If we can start to normalise data coming from all different sources we can attach functions to that. Maybe we’ll have functions market places that’ll automatically add AI, ML or image or video services…..I think it’ll be easier and easier for a developer to be able to use their stuff over time.”

As for too easy? One thing he is adamant about is that we are NOT moving to a “NoOps” world.

“The reality is you still have data, you still have data sources, you still have data bases, you still have infinite loops that you can write.” Constraints will be put in place by service providers, he says, as Iron.io did

“Over time you put more and more restrictions and guard rails in place so customers can achieve what they want to achieve but they can also not impact the rest of the systems…the idea of serverless is that you take as much of that worry away from the customer.”

And if that seems a way off, what about the blockchain world?: “I think there’s a worlds collide moment of serverless and blockchain…where we have marketplaces of smart contracts that can operate on distributed ledgers and marketplaces of serverless functions that can run on any cloud as a utility.”

But, he admits, “We’re a bit out from that.” But once we’re through Oracle’s Star Wars phase, presumably the sky’s the limit.

You can see Chad’s keynote at Serverless Computing London here: