The focus of Oracle OpenWorld was Oracle Cloud, and the company positioned Oracle's offering squarely against AWS, citing metrics showing their technology is orders of magnitude faster than Amazon's for database workloads. The statistics are impressive - compared to AWS, Oracle is said to perform "35X faster for online transaction processing (OLTP), and 1000+X faster for mixed workloads" - with the press release concluding that "the Oracle Cloud is optimized for running Oracle Database while Amazon Web Services (AWS) is not".
The CEO of AWS, Andy Jassy, had a short response to the performance comparison: "As far as I can tell, that was made up".
Oracle's core business today is still on-premises software. The fiscal results for 2016 show software licence sales, updates and support accounted for 70% of company revenue, while the cloud accounted for just 8%. Yet every one of OpenWorld 2016's keynotes was about the cloud, so Oracle are publicly pivoting to present themselves as a major cloud player. The Oracle Cloud has offerings in SaaS and PaaS but attention at OpenWorld was directed at IaaS:
Oracle today introduced the broadest array of Infrastructure-as-a-Service (IaaS) offerings in the industry, which include bare metal cloud servers that are 11.5X faster and 20 percent cheaper than the fastest solution offered by the competition.
One month before OpenWorld, Gartner produced their Magic Quadrant for Cloud Infrastructure as a Service (available in summary from AWS) which rated Amazon Web Services and Microsoft Azure as Leaders, and didn't include Oracle: "At the time of this evaluation, the Oracle Compute Cloud Service was not in general availability, and Oracle did not have another cloud IaaS offering in general availability that meets Gartner's definition of IaaS. Consequently, Oracle also does not have enough market share to qualify for inclusion."
InfoQ spoke to industry experts in the market leading clouds to ask their opinion of Oracle's future success in the public cloud space, and whether the focus on IaaS is the correct one. Mike Pfeiffer is a former member of the AWS team and authors Pluralsight Courses on AWS certification; Rik Hepworth is a Microsoft Azure MVP and regular conference speaker.
InfoQ: Oracle are focusing on IaaS in their Cloud Platform announcements. Is IaaS enough?
Mike Pfeiffer: It’s a good starting point for getting their existing customers into the cloud and I think they’ll have a bit of success in that regard. However, when it comes to attracting brand new customers, I think it will be tough to compete with the breadth of services offered by AWS and even Azure. The wider range of services offered by these other providers means customers have more options when trying to move complex applications to the cloud. The Oracle cloud platform will need to mature a bit more to compete with the likes of Amazon and Microsoft.
Rik Hepworth: I’d really like to say that no IaaS alone is not enough. A cloud provider that only offers IaaS is really just a hosting provider. That’s not because I dislike Oracle, but that I think cloud adoption is very much a journey that, whilst it may begin with IaaS migration, should move towards PaaS implementation of applications for any organisation. If there is no broad portfolio of cloud services to build those applications the customer stagnates in their IaaS environment and fails to drive value from the cloud.
InfoQ: What do you think Oracle needs to do to win a worthwhile share of the cloud market?
Hepworth: Differentiation is key to success for Oracle. They need to be able to convince potential customers that only they offer a robust and reliable service to host Java and Oracle database applications. If they can do that, the prevalence of Java in the enterprise could be a strong lever for Oracle to pull. The problem they face is convincing customers that they are able to provide a service that is as reliable as the competition. They will need to aggressively develop and deploy cloud services to show that they can build and deliver a service as expansive as those providers that have been in the cloud game for much longer.
Pfeiffer: They’re going to need more experience as a cloud provider. If you look at AWS, they’ve been doing this for ten years. Many of the products and services offered by AWS have been shaped based on the demands of their customers, some of which happen to be some of the biggest brands in the world. Obviously, experience is just something you can’t buy.
InfoQ: Microsoft can integrate Azure tightly with the .NET ecosystem. Oracle could potentially do the same with Java. Amazon don't have their own application platform, so why is AWS winning?
Pfeiffer: Again, I believe a lot of their success has to do with their experience as a cloud provider. AWS is winning because they’ve been innovating and solving customer problems longer than anyone else. Their platform is a reflection of that. The breadth and depth of AWS services gives customers numerous ways to solve problems, and that’s one of the keys to increasing adoption.
Hepworth: I think Amazon gained a lot in terms of first mover advantage. When Microsoft first presented the Azure cloud they were too far ahead of the customer base in their thinking. PaaS is absolutely what cloud adoption should be, but the lack of IaaS was a blocker for organisations who weren’t mature enough in their approach to cloud. Delivering a reliable IaaS service which acts as the gateway drug, then following up with a range of PaaS services to keep customers hooked as they migrate their applications to fully utilise cloud has worked well for Amazon. Where I think things are going to get very interesting pretty soon is the area of cloud-only services like Microsoft’s Cortana Intelligence suite. More organisations want to sweat their data and drive efficiency through clever automation and I see that area as the next big battleground.
InfoQ: With host-agnostic platforms like Docker, isn't there a new use-case for IaaS?
Hepworth: I really hope not! Now, before you get your pitchfork and torch to chase me down, let’s think about this. We’ve proven the benefits to organisations of PaaS over IaaS – lower management costs, easier scalability, etc. I think containers have their place, but to really drive adoption we need container services in the cloud that mean I don’t need to think about IaaS at all. I want to be able to treat my container deployments in the same way as I treat my serverless PaaS services – throw my containers into the cloud and know that they will run, and that the container service will scale according to my needs, but that I don’t need to think about anything beneath my container in terms of management. Without that approach, I don’t think the benefits of containers in the cloud can be fully realised.
Pfeiffer: Sure, containerized applications open up a lot of possibilities. But many people are still just in the initial learning stages with this technology. Many enterprises are still running big bloated applications that have tightly coupled components that don’t lend themselves well to some of these new patterns and practices. What we’re seeing right now is progressive businesses building high scale applications with containers, but it will take many other businesses a really long time to get to that point. I think this is especially true for enterprises. Initially I think we’ll see a lot of these legacy enterprise workloads moved over to a traditional IaaS implementation, and from there they may move on to embracing containers, or even moving their applications into some type of PaaS solution.