The Leverage of Bare Metal for Oracle Cloud

On year 2016, Oracle announced its Bare Metal Cloud as the company’s next-generation offering. Later Bare Metal Cloud got renamed into Oracle Cloud Infrastructure (OCI), to be differentiated from Oracle’s existing 1st generation of cloud service offering, which is called Oracle Cloud Infrastructure Classic (OCI-C). Since then, OCI is under heavy development by talents acquired from other top cloud service companies like AWS, Microsoft Azure and Google Cloud.

Without further delay, other cloud giants soon came up with their bare metal offering:

  • Microsoft Azure offers preview of its bare metal cloud on Nov 2017 [1]
  • AWS launches bare metal instances at AWS re:INVENT 2017 on Nov 2017 [2]


What is bare metal cloud?

According to Wikipedia, a ‘bare-metal server’ is a computer server that is a ‘single-tenant physical server’. The term is used nowadays to distinguish it from modern forms of virtualization and cloud hosting. [3] If you recall the virtualization technology like hypervisor or container taught in the 1st lecture, it’s simply a dedicated machine without virtualization. As you may know, virtualization brings in vertical layering for isolation, like VMs or containers and horizontal layering for management, like hypervisor or host OS. Those layering causes non-negligible amount of overhead and can potentially harm the neighbors during peak time. One selling point of bare metal is to address that. Isolation is guaranteed given by ‘single-tenant’.


How is bare metal cloud’s technical performance?

Accenture reports OCI is 5 times cheaper than other leading cloud solutions with same performance. [4] OCI also comes with huge delay in latency for its storage performance. [5]

On 2016 Oracle Open World, CTO Larry Ellison claimed the bare metal cloud offering can be 11.5 times faster in performance and 20% cheaper in price.

Oracle CTO Larry Ellison Keynote: AWS and Oracle Bare Metal Cloud Comparison

What about market performance?

In Q3, Oracle reported a 32% up in cloud revenue, which boosted ORCL stock price by 10% around that time. [6] For your reference, AWS remains a quarterly revenue growth of 40+% since year 2016 and even faster growth in its earlier stage. [8] Then we come to Q4 Oracle financial report, cloud revenue was hidden. [7] It led to a 12.5% stock price downhill at lowest around mid-June, 2018. I can’t say for sure the hidden report means bad performance in growth, but it’s highly likely. One thing to point out is the cloud revenue includes earnings from all SaaS, PaaS and IaaS. But given Oracle’s strength has been SaaS(HCM, CRM etc) and PaaS(Database as a Service) for a long time, the revenue share of OCI could potentially be even smaller.


What are the other elements when it comes to IaaS picking?

So, if bare-metal cloud is indeed a technology innovation and Oracle started earlier in field, what is other factors hindering OCI’s competition in the IaaS market?

First, cloud eco-system. There are many categories of IaaS customers. One set is the SaaS or PaaS bundled customers. At big data times, customers need not only compute and storage, but also smart solutions like AI, machine learning. While Google is all-in in machine learning services for a break through, Oracle still uses its well developed SaaS and PaaS to play to its advantage. It doesn’t mean Oracle’s market is small or shrinking, but more effort in the cloud eco-system could for sure help Oracle cloud, let along OCI.

Second, market demands for bare-metal. Bare-metal instance is mostly designed for enterprise customers. A typical use case is VMware. [9] But so far, most cloud service customers are satisfied with VM-based cheap commodity instances, for example Netflix and SnapChat are both building on top of AWS. Given VM-based instances is cheap and scalable, compensation between performance and price can be very dynamic. A possible outcome is a hybrid of VM-based instances and bare-metal instances, but not radically replace VM-based instances.

Third, sales and marketing strategies. Given AWS has being around for so long and built up its reputation, it becomes a no-brainer choice for lots of people. But bare-metal cloud is indeed a technology innovation and Oracle started earlier in field. If Oracle advertised it strategically with appealing lower price, there is a good chance to attract enterprise customers who could better perform with bare-metal instances. But given AWS and other cloud providers are already deep into the low-price battle, offering expensive infra with same price is not sustainable, unless Oracle found ways to compensate cost for its comparably expensive settings.















Users who have LIKED this post:

  • avatar

One comment on “The Leverage of Bare Metal for Oracle Cloud”

  1. This is a fascinating read, Hailun!
    Can’t believe I didn’t know about OCI until today (and man I wish it wasn’t renamed. Bare Metal Cloud would sound so unconventional, perhaps could be more of an attention grabber??)

    Now, the performance — it’s a beast!

    This is such a great case study of the tricky-ness of product-market fit. Clearly, OCI had a proprietary advantage with regard to its technology; what was not clear is who the target segment would be (and why they would be foolish not to jump on this product…)

    Now, I am very curious how the story of 18C would unfold… [1] Again, awesomely impressive technology. Unlike OCI though, I can see 18C hitting more technological and organizational pain points in enterprises… Would love to hear your insights on this.



Comments are closed.