Oracle Corp.’s future is on the line, but its co-founder, Chairman and Chief Technology Officer Larry Ellison seems to be denying the obvious.

What’s obvious is that Oracle is losing the cloud wars. What’s less obvious is that it’s in dire danger of losing its grip on the enterprise database market, which is, of course, where it achieved dominance years ago and which it continues to leverage as its core cash cow.

And it remains to be seen whether its deepening focus on enterprise software-as-a-service applications can pick up the growth potential it’s been missing in the infrastructure-as-a-service and platform-as-a-service segments. Just today, Oracle suffered yet another setback: It lost a bid to open up the Pentagon’s $10 billion JEDI cloud computing contract to multiple bidders, as the U.S. Government Accountability Office dismissed the complaint.

Oracle’s long decline in IaaS and PaaS

The most fundamental trend is that Amazon.com Inc. is surging, due both to nonstop growth in both its core e-commerce business and its cloud-computing unit. Amazon passed Google LLC parent Alphabet Inc. earlier this year to become the second most valuable publicly traded company in the world.

The runaway growth of Amazon’s cloud-computing business unit has come at the expense the enterprise database incumbents with which it competes in a growing range of data, analytics, integration and other cloud-centric markets. Amazon Web Services Inc. has grown steadily since it launched its core computing and storage services in 2006. In July, AWS reported nearly 49 percent revenue growth year-over-year for the latest quarter, which was up slightly from the growth rate it had reported in the prior quarter.

In the public cloud arena, the AWS-Oracle rivalry began in earnest in 2014. That’s when Amazon released Aurora and began migrating Oracle’s on-premises customers and their data, applications and workloads over to its public cloud. The migration away from Oracle Database has picked steam in the past several years.

To see how fundamentally Oracle’s back is against the wall in the IaaS and PaaS segments of the public cloud, consider these facts:

• Oracle is not even within shouting distance of the top three IaaS and PaaS public cloud providers by market share. According to the RightScale 2018 State of the CloudReport, only 5 percent of enterprises have adopted Oracle Cloud, compared with 68 percent for AWS, 58 percent for Microsoft Azure and 19 percent for Google Cloud Platform.
• Oracle continues to show disappointing growth in public cloud revenues and has struggled to build its seven-year-old public cloud offering through a combination of strategic acquisitions and organic development.
• Oracle has failed to pick up notable market share in cloud infrastructure. AWS has a commanding lead in that market and is followed by Microsoft Corp., Google, Alibaba Group Holding and IBM Corp.
• Oracle recently reduced investors’ visibility into the financial health of its own public cloud offerings when it stopped disclosing the amount of revenue it brings in from that business.
• Oracle’s window of opportunity for bootstrapping itself into the top tier of public cloud providers through strategic acquisitions has effectively passed.
• Oracle principal enterprise cloud enhancements that were announced at OpenWorld were underwhelming, amounting to little more than the table stakes needed to stay competitive in public cloud, but not offering any slam-dunk differentiators.

Oracle’s best hope: Ride SaaS

Of course, Oracle has been on a tear in the SaaS side of the public cloud. As I pointed out during OpenWorld, Oracle’s strongest bet in the public cloud arena will be to grow its already impressive enterprise SaaS application portfolio, though even there it remains behind the market-share leaders SAP SE and Salesforce.com Inc. in enterprise resource planning and customer relationship management, respectively.

At the event, Oracle Chief Executive Mark Hurd made several predictions that support the pillars driving many of its announcements at OpenWorld. Hurd predicted that by 2025:

• AI will be integral to 100 percent of cloud applications.
• Customer interactions will be automated 85 percent of the time.
• Digital transformation will result in 60 percent of all IT jobs being in brand-new categories, such as supervisors for robots, smart-city technology designers and AI-assisted healthcare technicians.
• Blockchain will be integral to trusted information exchange in “virtually all applications.”

With those trends in mind, Oracle’s announcements then provide a clear foundation for deepening its cloud application differentiators. Wikibon is impressed with the breadth and sophistication of the new enterprise digital assistant capabilities embedded in Oracle Cloud’s enterprise resource planning, human resources, customer relationship management and customer experience applications, supporting AI-driven predictive, prescriptive, personalized and contextual decision guidance, what Oracle refers to as “intelligent process automation.” Likewise, Oracle announced several new enterprise-ready blockchain applications for supply chain management.

In the enterprise cloud SaaS application market, Oracle is far and away more advanced than its rivals in embedding AI-driven contextual guidance in its solutions. In the coming year, Wikibon recommends that Oracle deepen and extend that SaaS functionality in several ways:

• Incorporate its AI digital assistants into robotic process automation solutions that enable knowledge workers to build intelligent software robots that drive a wider range of administrative functions that are supported out of the box in its Fusion Cloud Applications;
• Expand the range of out-of-the-box enterprise resource planning, customer relationship management, human capital management and supply chain management decision scenarios supported by Oracle-built and -trained AI within its cloud-based digital assistants;
• Incorporate Oracle’s blockchain platform its AI DevOps tooling, perhaps to provide an immutable log for end-to-end transparency of data preparation, modeling, training and serving steps for governance and compliance purposes;
• Launch use-case focused stacks of the newly announced Oracle Linux Cloud Native Environment to help bootstrap its customers’ development of containerized and orchestrated cloud-native AI digital-assistant microservices for public, private, hybrid, edge and multicloud deployments;
• Extend its AI digital assistants to its Oracle IoT Cloud for Industry Solutionportfolio to drive prescriptive guidance into a full range of edge devices across industries and every business function;
• Provide partners and customers with a rich data science toolchain workbench, leveraging its recent DataScience.com acquisition, to support development, training and deployment of AI to address more complex, specialized and bespoke business requirements that would benefit from in-app digital assistants; and
• Allow customers to embed their own bespoke AI models in these applications to tune the recommendations to their specific enterprise resource planning, customer relationship management, human capital management and supply chain management requirements.

Enhancements such as these will be necessary for Oracle to address the full range of enterprise requirements for digital transformation. Doing so well would help the company pull away from the pack in the one segment, beyond enterprise relational databases, where it remains in the top tier in market share: enterprise-grade SaaS solutions.

Amazon’s accelerating migration from Oracle Database

Nevertheless, Oracle’s foothold in the IaaS and PaaS segments may continue to decline unless it radically revamps its go-to-market strategy. It may need to accept the fact that AWS, its core competitor, seems determined to end its operational dependence on Oracle’s flagship enterprise database.

Amazon has been an Oracle customer for most of the time since the Seattle-based e-commerce startup took root in the 1990s. Amazon has spent hundreds of millions of dollars on Oracle technology, including $60 million about a year ago. In last December’s earnings call, Larry Ellison told investors the Amazon account brought in $50 million for the quarter.

However, Amazon is far from a happy customer. For many years, it has been telegraphing its intention to wean its core e-commerce business operations away from dependence on Oracle Database in favor of AWS’ own equivalent offerings. Since 2014, Amazon has been migrating its e-commerce business away from Oracle Database and onto native AWS cloud databases, especially Redshift, Aurora and DynamoDB.

Amazon is reportedly less than two years away from migrating entirely away from Oracle Database in its operations, anticipating the process to be complete by the first quarter of 2020. The most recent milestone in this migration was when Amazon’s consumer business “turned off” its Oracle data warehouse on Nov. 1, having migrated the service entirely to Redshift.

By end of 2018, AWS expects that it will have 88 percent of its Oracle databases and 97 percent of critical databases moved to Aurora and Dynamo DB. AWS is also ramping up incentives and tools to make it easier for Oracle customers to migrate their data to AWS’ data stores and tools.

Examining Oracle’s defensive counterattack

What’s undeniable is that Amazon has delivered a one-two punch to Oracle’s solar plexus. In addition to stealing Larry Ellison’s oxygen supply in IaaS and PaaS, AWS has rapidly matured its diverse cloud data offerings into a formidable alternative to Oracle and every other legacy provider of enterprise databases.

Oracle’s usual approach to competitive challenges has always been to come out swinging. Ellison last month in his OpenWorld keynote boasted about Amazon’s dependence on Oracle technology. At OpenWorld, he engaged in his usual aggressive counterattack. He disparaged AWS’ Aurora relational database and Redshift data warehouse, arguing that they are inferior to Oracle’s equivalent data platforms. On many occasions, Ellison has claimed that AWS doesn’t have the technology to rid itself entirely of Oracle databases.

Ellison recently claimed that the public cloud isn’t safe enough for production enterprise applications, by which he was primarily implying AWS’ offering. He has even gone so far as claiming that there is a “fundamental problem” with the architecture of public cloud environments, a line of argument that appeared to undermine the value proposition of Oracle’s own public cloud offering.

In his OpenWorld keynote, Ellison mocked a recent incident that caused massive glitches for Amazon’s e-commerce unit. He referenced a CNBC news story about a 25-page Amazon “correction of error” report discussing detailed how a move from Oracle to its own Aurora cloud database resulted in database degradation, impacting a single Amazon retail fulfillment center.

CNBC reported that the snafu was related to a breakdown in an internal program called Sable, which is used by Amazon to provide storage services to its retail and digital businesses. An Amazon internal report cited in the CNBC story stated that “Oracle and Aurora PostgreSQL are two different [database] technologies” that handle “savepoints” differently, for “complex error recovery” in database applications and that, on Prime Day, an “excessive number of savepoints was created, and Amazon’s Aurora software wasn’t able to handle the pressure, slowing down the overall database performance.”

An Amazon spokesperson has said the issue with Aurora’s implementation in that fulfillment center had nothing to do with the Prime Day disruption. The latter incident was due to an issue with AWS DynamoDB on the retail website. AWS also said none of its other users had workloads impacted because of related issues in the fulfillment center impacted by the Aurora disruption.

In a tweet, Werner Vogels, Amazon’s vice president and chief technology officer, attached a detailed technical explanation for the disruption’s cause, resolution and impact. He said its impact was “delaying shipping of about 1 percent of packages [from only that fulfillment center] for a short period of time (unnoticeable to customers).” He added that “the problem was quickly diagnosed and completely resolved by simply removing the unnecessary savepoints that had been inadvertently left in the retail application. No changes were required in Aurora.”

Putting the issue with the single fulfillment center into broader perspective, Vogels wrote in that tweet, “Our Fulfillment Centers have migrated 92% of DBs from Oracle to Aurora with better avail, less bugs and patches, less troubleshooting, less hw cost.”

As Amazon’s database tech matures, Oracle’s shows its age

It’s not at all clear that Ellison’s counterpunching is hitting its intended mark. Over the past several years, AWS and its cloud data services — such as Redshift, Aurora and DynamoDB — have become as dominant in the public cloud area as Oracle Database ever was in the enterprise data center.

AWS’ maturation of its data services stems in great part from its parent’s longstanding practice of eating its own dog food. Since Amazon’s founding, it has been developing software for its internal applications and then, as those prove themselves out in production environments, turning them into products in AWS’ public cloud for the benefit of paying customers.

Oracle’s database has not met the test of cloud hyperscaling that AWS demands. In fact, Amazon has struggled to scale Oracle’s database fast enough to meet its customer demand. Recognizing the limitations of Oracle’s database, AWS stopped developing new technology around that data platform years ago. At last year’s AWS re:Invent conference, Jassy challenged partners to “find a lot of customers using Oracle who are actually happy about it.”

Recognizing an opportunity to encourage defections from Oracle’s huge database customer roster, AWS now offers a tool that allows businesses to move Oracle databases to its cloud. The Database Migration Service, which supports Oracle’s software, has handled the transfer of more than 80,000 databases to AWS as of July.

AWS has always been adept at commercializing its hard-won operational expertise into tools and solutions for the broader enterprise IT market. If Amazon can leverage its operational lessons learned on its own migration away from Oracle Database and turn it into a high-quality migration service for AWS customers, it might be able to encourage more customers to switch from Oracle. But that’s a big “if.”

Database migrations are always hard work, and they tend to take a lot of time, money and technical expertise. This reality is the reason why so many Oracle database customers feel locked in and it’s why a migration tool is rarely a quick-fix, rapid-value proposition. If Oracle can’t rapidly stoke growth in its database and cloud offerings, AWS may very well step up its development of migration offerings, partnerships and tooling to encourage further migrations away from Oracle Database.

Perhaps Oracle can fight back through a hybrid cloud strategy under which its SaaS offerings are so intimately linked to its enterprise database that customers have a real functional advantage for continuing to deploy the DBMS on-premises. AWS has no countervailing SaaS strategy that can hope to contend with Oracle, SAP, Salesforce and others in that high-margin side of the cloud business.

How does AWS plan to make further inroads into the enterprise database market shares of Oracle, Microsoft, IBM and other longtime tech companies? Can AWS push back against Oracle, SAP and Salesforce’s dominance in SaaS?

Come see what AWS and other executives have to say on theCUBE live at AWS re:Invent 2018, from Tuesday through Thursday, Nov. 27-29.

Image: Håkan Dahlström/Flickr