One of the benefits of being a smaller analyst firm is that you tend to see trends more quickly, and because you don’t have layers of bureaucracy, you can start analyzing and communicate on those trends quickly. One of the things that Moor Insights & Strategy has been talking about since before its popularization is the notion of a hybrid cloud with one control plane and maybe a data plane across the enterprise’s public and private cloud.
The one related prediction we got wrong, however, was that OpenStack would drive it. Oh well, you can’t get it right, every time. Years ago, when we talked about hybrid cloud, the public cloud players would scoff. And, of course, the on-premises systems and software providers thought it was a great idea because it had everything to gain. Oh, and it is an option enterprises still want today. The argument on the viability of the hybrid cloud has been put to rest, thankfully, with the announcement from the public cloud leaders with Google Anthos, AWS Outposts, Azure Stack, and IBM Cloud Private.
The public cloud is nearly fifteen years old and still only 20% of enterprise applications are in the public cloud and that also means that 80% of applications are still on-premises. There are incredibly good reasons for on-premises, including latency, data residency, data sovereignty, control, security, and just not enough time for enough time or resources to take applications to the cloud. Oracle recently announced an interesting and unique hybrid cloud offering, called Oracle Dedicated Region Cloud@Customer that I think promises a strong value proposition.
The concept is simple and that is to take everything in the Oracle public cloud and duplicate it on-premises. This includes all infrastructure, operating system, tools, security, databases and Oracle applications and makes those available on premise behind an enterprise’s firewall, fully managed by Oracle.
Historically, I have not been very bullish on the Oracle Cloud efforts over the many years. Oracle’s strategy went different directions and believe in many ways, set itself up to fail early on. I do not believe it was ever a promising idea to try to compete head to head against AWS on commodity workloads as this required scale and vertical integration the company wasn’t willing to bet. But I always saw the opportunity for the Oracle Cloud for current customers if it could optimize its offerings for its customers which are far and wide, most of the Fortune 1000 and watch its pricing. So, Oracle had to produce something that was differentiated and first to market, which I believe it promises with Oracle Dedicated Region Cloud@Customer.
As I said before, the concept of the private, dedicated region done right sounds quite simple. You take every capability that’s currently in the Oracle public cloud, all the hardware and all software and you make that available behind a cage in the customer’s data center, including database and apps. The customer doesn’t have to touch any of the infrastructure, it is fully managed, but customers but can manage all the capabilities behind a single console, be it public or private cloud.
Specifically, Oracle is offering:
- Compute: 6 flavors
- Storage: 6 flavors
- Management/governance: 9 flavors
- Security/identity/compliance: 5 flavors
- Network/edge/connectivity: 6 flavors
- Appdev: 13 flavors
- Data management: 12 flavors including
- Analytics/big data: 8 flavors
- Oracle Fusion SaaS apps: 5 flavors
As you can see, this is a full-flavored private cloud.
Early customers like Nomura Research Institute (NRI) and Oman ICT Group are exactly the types of who you would expect. These enterprises are large Oracle customers and have special data residency and sovereignty requirements but who want the benefits of a cloud environment thats data never leaves the on-premises data center. Having data never leaving the on-premise data center makes it easier and in the case of Oman, allow it to operate in a regulated environment.
Other things I thought were strong about the offering is its simplicity. Pricing is the same as you would pay in the public cloud, starting at $6M per year, and can be used against basic IaaS, PaaS and SaaS products. The SLAs (99.5% uptime, performance, API error rate) which are especially important for those applications that are vital to the lifeblood of a large business or a government are the same as well. Compliance certifications like SOC 1, SOC 2, ISO 27001, ISO 27017, (NIST) 800-53 rev4 publication are the same as well. Like I said, an exact copy, fully managed by Oracle.
Competitively, Oracle is first to market with this kind of capability. I do believe that AWS will offer this this broad, “connected only if you want it capability” over time. While AWS doesn’t advertise SLAs, it does negotiate these on a customer by customer basis. After AWS I see Azure then Google offer this over time. But Oracle can bask in the glory now.
I believe one thing that enterprises need to consider is Oracle’s cost competitiveness over the next five to 10 years. Oracle appears to be priced competitively for long term commitments at this moment, but will it make the investment to ride the cost curves down? I am impressed that it has its own homegrown processors to defray costs and optimized for scale-up applications. But will it continue development of those, and will it be developing more homegrown infrastructure to reduce the cost of machine learning, training and inference as well? I think those are important considerations to think about as enterprises make commitments that are easily five to 10 years in duration. Make no mistake, no customer is going to go through the six-month process to install the equipment on-prem for just a three-year minimum commitment. To be clear, I am under no illusion that cost is the only element as reliability, performance and security for many workloads are more important attributes.
As I was a believer in the hybrid cloud a decade ago, I am a believer in multi cloud, which gets made fun of by many public cloud vendors today except for Google Cloud, particularly those who would lose business, if more people truly adopted a multi cloud model. Therefore, it’s important for customers to consider how Oracle plays nicely with AWS, Azure, Google and the IBM Cloud. If you believe that multi cloud is the future, then it’s important that any cloud provider you work with interoperates at some level with all the public cloud vendors you would consider working with.
While I have been less than inspired by previous Oracle cloud offerings, I think Oracle deserves many kudos for its Dedicated Region Cloud@Customer. Conceptually, this is exactly what enterprises want, and that is to have the exact same, broad capabilities and pricing in the public cloud in a fully managed private cloud, paid for in a consumption model and managed from a single pane of glass. I believe the value proposition will resonate with large, regulated enterprises with mission-critical, higher availability applications running on an Oracle database or those running Fusion apps. Its vitally important for Oracle to communicate how it will remain cost and features competitive over the long haul as these decisions are long-term decisions. I believe AWS, Azure, IBM and maybe GCP will ultimately offer similar capabilities but until then, Oracle gets to bask in the sunlight.