Oracle’s PaaS/IaaS progress - becoming real

Profile picture for user brianssommer By Brian Sommer April 19, 2017
Are Oracle's IaaS and PaaS solutions real? Here’s what Brian Sommer learned at a recent analyst event centered on just such a question. The reality seems to be a lot of progress, but with some lingering concerns.

2017 blog - ORCL IaaS pic 2
Oracle's going big on IaaS and PaaS solutions, but are they real? It's not a new question, but it remains a valid one that deserves attention - which it got at a recent analyst event staged by the vendor.

Flash back to late 2015 and I saw Oracle executives define a new competitive landscape. They described a world where the applications, platform and infrastructure spaces were getting upended. Old competitors would cede market share to newer players. The new winners were companies whose solutions were designed for a more cloud-y world. I wrote of this in this 2015 diginomica piece.

At the same event, I saw Larry Ellison, CTO and Founder of Oracle, demonstrate a number of cloud capabilities his firm was developing to make Oracle more relevant in this changing business world. He then showed how easily he could transfer a customer’s on-premises applications, database and more to a cloud computing solution Oracle had developed. The key points I recall from that demonstration were:

  • Oracle wasn’t just selling Software-as-a-Service (SaaS) applications like Fusion. It was now marketing Infrastructure-as-a-Service (IaaS) and Platform-as-a-Service (PaaS) capabilities.
  • On-premises to cloud (or vice-versa) migrations would occur relatively easy.
  • Oracle was developing a competitive offering to Amazon AWS that was targeted for larger enterprises.

What Oracle was announcing then was a strategy to retain its customers. Oracle had noticed existing customers moving some of their computing workload and development efforts to third-party cloud infrastructures instead of expanding their own data centers or private clouds. Those moves might also tempt users to use other systems software tools, including open source technologies, which would further erode Oracle wallet share.

Coinciding with these trends were the rise of platforms from application software vendors (e.g., Salesforce and Workday) and the aggressive movement of Microsoft as a platform and infrastructure provider to other application software firms.

Ellison’s initial OpenWorld demonstration was in the Fall of 2015. Diginomica’s Dennis Howlett was also there and he noted that:

Customers can move seamlessly between on-premises, Oracle cloud, somebody else’s cloud and program using pretty much any language they want including nods in the direction of fashionable Docker containers, node.js and the rest said Ellison. But is that real?

So, as 2015 came to an end, we saw Oracle wanting to take on new competitors and face new markets with a focus on the utility computing space. The question Dennis asked then is still valid - is this real?


The first clues came in the Fall of 2016. Ellison doubled-down on the importance of IaaS and PaaS to Oracle and its customers at the following OpenWorld. Oracle was a building a public cloud solution to compete with those of Amazon and Google. Diginomica’s Derek DuPreez covered the announcements at that event. He reported:

We saw on the opening night keynote how now CTO Larry Ellison declare that “Amazon’s lead is over”. A bold claim considering that Oracle’s IaaS revenues pale in comparison to that of AWS. However, Ellison’s statement falls on the claim that Oracle’s IaaS offering is not only cheaper than AWS, but it offers twice the cores, twice as much memory, and four times as much storage.

The response to this has been fairly critical, considering AWS’s multi-year lead on Oracle. I questioned why Oracle would want to put so much focus – particularly at its main event of the year – on what is essentially a low margin, commodity business?

But having had discussions with people over the past 24 hours, the focus on IaaS for Oracle is important because currently its customers that are shifting workloads to the cloud are doing so with Amazon. If that happens, it then struggles to maintain a PaaS conversation with those customers, as well as other cloud discussions further up the stack.

Today's reality

Recently, Oracle invited a few industry analysts to understand the progress they’ve made in these pursuits. Are these new IaaS and PaaS solutions real? Here’s what we learned at the recent analyst event:

  • The team Oracle has assembled for its IaaS and PaaS offerings in Seattle is big, growing and experienced. Not surprising, many of these team members are ex-Amazon AWS (some from Microsoft Azure Cloud) and bring with them the knowledge of what did/didn’t work with the first movers in the space.
  • Customers can get anything from a bare-metal solution (where the customer simply ports their own virtualized solution to the Oracle Cloud), to a public-cloud friendly environment for running Oracle applications, to a private cloud-in-a-box environment. Several other combinations are possible, too. One of Oracle’s offerings, the cloud-in-a-box, is a private cloud, on its own machine, preconfigured and delivered to a customer’s data center. The customer can easily move their on-premises custom applications and packaged Oracle applications (e.g., PeopleSoft) to this device. The customer gets a behind the firewall solution lifted and shifted with minimal fuss and priced on a subscription basis.
  • The team wasn’t as hyperbolic as the messaging was at recent OpenWorld events. Why? The team can point to actual deliveries of a number of enterprise-level (not consumer, small-business, startup or hobbyist/citizen developer) capabilities to the IaaS and PaaS environments. From their perspective, the kit for an ‘enterprise cloud’ solution includes: uncompromised security, top performance, high reliability, intuitive/simple pricing and supportive of Oracle’s applications.
  • Oracle appears to have signed a number of customers to its IaaS and PaaS offerings. Market traction is growing for these solutions. Oracle reported that their PaaS solution already has over 11,000 customers.
  • Oracle claims some 2,900 applications available in their platform marketplace.
  • Oracle’s cloud-in-a-box offering also lets Oracle deal with data sovereignty issues as it can drop an operational device most anywhere in the world. However, the devices are only shipping to a select number of geographies currently. Oracle states that these devices are selling briskly.
  • Oracle says the hardware in the cloud-in-a-box (Oracle’s name for this is: Cloud@Customer) will be refreshed every four years. The minimum term is 1 year (with an additional 1 year early termination fee).
  • Oracle will use its own employees to service these devices in many areas and will utilize partner firms in some geographies. That process is still being developed.
  • The way these IaaS and PaaS technologies can be deployed is interesting as Oracle can deliver capabilities at the country, area or company level for customers.
  • Oracle is also marketing DBaaS (Database-as-a-service). It’s an interesting idea as Oracle is bundling its RDBMS, in-memory technology, security and other features with other data tools (e.g., Hadoop, NoSQL, Spark). This is to aid companies and developers with accessing different types of data and processing workloads.
  • Pricing comparisons were offered up at several points during the meeting and being a low (or the lowest) cost provider is a point of pride for this team. Individual customers will need to do their own price comparisons based on their particular needs and timeframes. More importantly, we were shown actual prices on price schedules – this pricing transparency almost never happens with application software.

So far, so...

Away from that list of claims, there are some areas that warrant additional discussion. For example, the marketing of these solutions was not clear. While this was a technical briefing with a technical team, it appears that Oracle will have different levels of penetration and marketing ease selling into some segments.

Some companies that are faced with obsolescing computer hardware might choose the cloud-in-a-box option if they are adamant about not going public cloud. These firms want a private cloud experience without the big CapEx costs. These customers are likely laggard technology buyers who will cause Oracle to incur high sales costs and possibly lower initial margins.

A larger group of existing Oracle customers would likely want the ability to lift and shift their existing applications to Oracle’s public cloud. Oracle would then be on the hook to maintain the technical environment and scale/usage issues would be Oracle’s concern not the customers’ concern. As before, both prospective customer segments get a technology refresh. This group of prospective buyers wouldn’t necessarily be laggards.

No, many of these will be middle-of-the-road Oracle application software customers who want to get out of the data center business. Porting these customers to Oracle’s (or Amazon’s, Microsoft’s or Google’s) cloud could help these customers avoid big CapEx bills when it’s time for a hardware refresh.

Oh, and let’s not forget how big this potential lift and shift market for Oracle could be. This could include a fair number of existing Oracle customers with Oracle EBS, PeopleSoft, JDEdwards and other product lines. A recent Oracle fact sheet pegs their application customer count at 110,000. That segment alone could keep them busy for the foreseeable future.

Whether the customer lifts and shifts to the private cloud-in-a-box or the Oracle public cloud solutions, the customer will have access to the newest Oracle RDBMS and other tools. Customers using older Oracle RDBMS versions should check with Oracle to understand what, if anything, they’ll need to do to take full advantage of the lift and shift.

Some readers will no doubt note that customers could go to a competing cloud environment via VMWare or other virtualization technology and stay on an older version of the Oracle RDBMS. That’s true and running virtualized environments on Oracle’s cloud is also possible. Should Oracle users want Oracle to provide more that brute-force computing power (e.g., automatic patching of systems software, advanced database security, etc.), then they’ll want to run a newer version of the Oracle solution.

Oracle executives also indicated that a growing market segment for their IaaS/PaaS solutions was in new software developers – especially those doing large-scale computational work. For example, computer graphics engineers at movie studios may want access to large, fast computing power for rendering graphics for movies. Others may need discrete chunks of computing power to process specialized analytics. These firms want cloud solutions that are extremely fast, highly scalable and capable of processing massive datasets.

It’s that last segment that really put a sharp point on things for me. This type of customer wants POWER and they want it only when needed at an affordable price point. These buyers are a type of enterprise buyer that will cause traditional technology firms to either change their culture and pricing approaches or face extinction.

That last point triggered the really big observation of the day. Oracle’s pricing and terms for these IaaS and PaaS solutions appeared to be affordable and flexible. Granted, I didn’t get a contract to take home and review, but, the terms floated before us were quite different from those seen in application software deals (cloud or on-premises).

Market forces are dictating that cloud infrastructure players must be flexible and inexpensive regarding their service or customers will look for true love elsewhere. Firms that are competing with the likes of Amazon have to understand the effect of 50 or so consecutive price reductions that Amazon AWS has put in place over the years and how this action changes the attitudes of new buyers. Utility computing is a commodity. Scale, positive customer experiences and transparent pricing are market requirements.

Application software firms that try to bring their opaque contracts and practices into this space will likely be unsuccessful. Unlike the decades-long lock-in that comes with many application software products, utility computing can easily be a month-to-month environment with really low switching costs.

So, how will Oracle have one set of pricing terms, contracts, business practices, etc. for its applications business and a very different kind for its IaaS and PaaS business? That’s got to be a great topic at their sales and board meetings.

This leads us to another concern: will Oracle’s past business practices impede its cloud efforts? Oracle was listed as the fourth most frequent auditor of customers’ software usage according to Net(net). Pricing of their application software on non-Oracle clouds was also flagged by Net(net) as being twice as expensive as when it runs on Oracle’s cloud (see “Oracle Doubles the Cost of Cloud Deployments”). If true, then recent technical gains made by Oracle’s team seem impaired by the gaming of its application pricing.

Finally, will other ERP vendors adopt the Oracle IaaS/PaaS solutions? Most application software firms have already made their cloud and platform choices. SAP has HANA. Infor uses AWS. Workday has its own software stack that contains a lot of open source tools. A number of other firms have chosen to use aspects of Microsoft’s stack. This market segment may be a tough one for Oracle to win in the short-term but one that it should continue to pursue for the long-term.

My take

Oracle’s made a lot of progress technically and in building out capability. They’ve also acquired a number of customers for these new solutions, too. The addressable market for them could be big and it will take time for it to mature. Soft issues (e.g., branding, marketing and pricing) may need additional sorting out in the near-term especially as the company melds a low-cost commodity utility solution with a high-touch applications business.