Oracle stays the course on IaaS

In its ambition to be the premium cloud of choice for enterprises, Oracle has invested in datacenters, services and features over the past year. Its big challenge lies in convincing the market of its credentials and its value, but it has made progress.

When Oracle launched its cloud in 2016, it faced an uphill struggle. We think there is a market for a premium cloud, but does Oracle have the credibility, and perhaps the appetite, to be that premium option? The fact that it has been consistently adding new features, services and regions over the past year shows that it's not just offering databases alongside virtual machines – it is vying to be a competitor to AWS. But its challenge continues to be showing potential customers that it is the better option, both in terms of quality and price. AWS and Azure are go-to brands – buyers don't need convincing to give them a shot. Oracle, at least for the time being, needs to throw effort into working with enterprises to show its value. It needs to be flexible, generous and attentive with prospects (way more than its competition), and it needs to provide that level of service to all prospects, not just the conglomerates it is used to working with. If this means new commission plans and incentives are needed, so be it. Oracle is investing heavily in its cloud; it needs to invest heavily in its brand and its engagement to really win hearts and minds.

Context

Few would argue that Oracle came early to the cloud game. It launched a public cloud compute and storage service in 2015, but had few differentiators compared with the behemoths of AWS and Microsoft. In September 2016, Oracle launched a new cloud, Bare Metal Cloud, which has now been renamed Oracle Cloud Infrastructure (OCI).

The premise was a simple one: The likes of AWS had pioneered cloud computing, but Oracle would take the lessons learned from those efforts to make a better-performing option for enterprises. Security, performance, off-box network virtualization, SLAs and the resolution of noisy-neighbor problems would be built-in from the outset.

It hired some of those pioneers. Senior VP of engineering Don Johnson (one of AWS's original architects) was hired in 2014, and the company has hired more than 1,000 people from other cloud firms. Its cloud-development function is headquartered in Seattle, where AWS, Microsoft and Google all have their cloud HQs, and where cloud talent is easier to come by.

We covered its ambitions in January, and identified the key factors that would determine the company's success or failure: Do enterprises want a premium cloud service? Does Oracle have the credentials to become the cloud service of choice for enterprises? The analyst summit at Oracle OpenWorld in San Francisco provided an update.

A busy year

Oracle reports its PaaS and IaaS revenue together, which makes understanding its IaaS growth difficult. FY16 to FY17 revenue increased from $0.9bn to $1.4bn, equivalent to 60% YoY growth. The company claims to have added 14,000 IaaS and PaaS customers to OCI since its inception, almost all of them existing customers of its licensed software. Oracle's overall revenue in 2016 was $37bn, so IaaS and PaaS still represent a small slice of the pie.

A challenge Oracle faced from the beginning was its tardiness to the market. Sure, it could copy and perhaps improve upon existing public cloud offerings, but it would have to do it faster than the rest of the market. AWS, for example, has over 70 services, so there is a lot of ground to cover. Over the past year, Oracle has released 50 services and features – starting from bare-metal compute and storage, the company has added virtual machines, databases, database clustering, load balancers, audit capability, compliance, monitoring, logging, authentication and new images. From a single datacenter in Phoenix, it has expanded to Ashburn, Virginia, and Frankfurt; it is targeting London for early 2018 and APAC further down the line. It has also released and open-sourced a new serverless capability called Fn and a Docker-native platform called Fn Flow for composing serverless applications. The company hopes to distinguish its serverless offering by making it cloud-agnostic, although Java is first among equals in terms of supported languages. Oracle realizes that its capability isn't as broad as AWS's, but its rate of development shows it can achieve a lot in a short amount of time.

Oracle has identified three areas where it has had particular success: enterprise lift-and-shift migration, particularly with regard to its databases; infrastructure-heavy workloads that require large bare-metal and virtual machines and SSD storage; and cloud-native. Its Oracle Cloud at Customer service allows the same IaaS and PaaS to be consumed on-premises with no difference in price, although a commitment is required.

During the event, Oracle announced new SLAs on full-stack availability, including the application, the availability of the API, the availability of its Exadata servers and even performance (including IOPS and throughput). There is even an SLA on hardware repair. Oracle must be confident of its performance characteristics if it is willing to place a guarantee, although we will have to wait for launch before understanding its potential. The company recently initiated a new flexible pricing model too, specifically targeting AWS.

Oracle appears to have real ambition, and is executing its strategy in a bid to catch its competitors. But do enterprises want a premium cloud?

Premium value, commodity price

Oracle wants to be more than just the best cloud – it wants to be a premium cloud at commodity cloud pricing. It claims that, through a better architected and better managed public cloud, it can offer the best performance and some of the lowest prices in the market. Of course, the cost of cloud depends on a huge number of variables, and we can't say that one cloud is always the cheapest, but Oracle's pricing does seem to be competitive with the hyperscalers it is looking to take on. The Cloud Price Index has found that, over the past year, companies with the lowest prices have not necessarily derived more market share, and vice-versa. This suggests that, although buyers want to pay a price that is reasonable, they are willing to pay more for differentiators. Oracle is targeting both plays: value and cost.

In 451 Research's Voice of the Enterprise: Hosting & Cloud Managed Services, Organizational Dynamics 2017 report, 44% of 515 respondents stated that they would pay a premium for an enhanced SLA on performance/uptime; 34% stated that they would pay a premium for enhanced customer support. The median premium for these enhancements was about 20%. Buyers see value in services way beyond just the basics. The challenge for Oracle is convincing customers that it offers the best capability for the best price – there are others in the market with stronger credentials and reputations.

At the analyst summit, we asked a number of customers how they ended up using Oracle in a market dominated by more cloud-established players.

A SaaS provider of building-industry project management software recently transitioned its infrastructure from IBM Bluemix to Oracle, following a chance meeting with a sales contact in January. The customer was not an Oracle customer previously, and does not consume Oracle databases – the key decision criterion was performance. The customer used benchmarking software to compare performance across Oracle, AWS, Azure and Google, and stated that performance was significantly better using Oracle at a comparable (if not better) price than its competitors. There were still signs of the old-fashioned Oracle – it initially wanted an up-front payment (in contrast to the generous free credits in the range of $50,000-200,000 offered by the hyperscalers), but through a process of escalation in the sales organization, the customer was promoted to the Velocity program, which incubates and supports new potential clients.

In a discussion with an Oracle SI partner, the partner stated that Oracle's offering solved a lot of problems, giving it access to powerful cloud capability located in its native Germany. Again, there were signs of the old Oracle – the partner believed that Oracle's cloud was the best one for the job, but was frustrated by how difficult or expensive it is to consume Oracle licenses on other clouds. Earlier this year, Oracle effectively doubled the price of its AWS instances overnight.

Another security vendor customer cited Oracle's bare-metal base as its key decision criterion. It has a need to create virtual machines upon its hypervisor without nesting, which it couldn't do using AWS, Azure or Google. The customer was already an Oracle customer before moving to Oracle Cloud Infrastructure. We anticipate that most Oracle Cloud Infrastructure sales will be to existing Oracle customers.

Of course, at an Oracle Summit, we were more likely to meet those happy with Oracle's infrastructure than those who weren't. However, the candid feedback from those customers suggests that Oracle has made progress in its ambitions.

Competition

AWS has a huge first-mover advantage, and it is the de facto cloud platform – its strategy is to act as a huge, inexpensive toolbox, upon which developers can build powerful and cutting-edge scalable applications, charged in an on-demand fashion. Microsoft is primarily looking to differentiate on being the hybrid Windows environment of choice: It has public and private cloud capability (through Azure Stack), accessed using the same APIs, GUI and even pricing methods. With its hybrid-use model, it is incentivizing Windows users to use Azure for cloud. Google is trying to take on AWS, but its key messaging is based on being developer-friendly and fit for hyperscale – Google has created groundbreaking technology through its search and other activities, and its cloud is aimed at giving developers access to this capability for pennies. IBM Bluemix and Oracle have similar messaging – bare-metal, IaaS/PaaS/SaaS capability developed for enterprises, across hybrid cloud environments. Rackspace is looking to be a managed service provider, rather than a hyperscaler, differentiating on support and added value.