tl;dr = Oracle’s cloud strategy is 1) to allow customers to easily bring what they have at strong price points so 2) customers will build the future with Oracle. They’re continuing to execute on that strategy with financial + customer success.
For those of you following along at home, the first Oracle Ravello Blogger Day (#RBD1) coordinated by TechReckoning was May 24 last year. If you’ve somehow forgotten all that goodness, you can read Chris Wahl‘s vintage blog post.
It’s 2018 and time for #RBD2 – Oracle Ravello Blogger Day 2. You can just read the Twitter feed at the prior link or keep reading for a few thoughts.
After a great opening and recap by John Troyer, Simon Law kicked off an agenda review.
Oracle’s overaching message through all the sections was that they believe they can be the best environment for “lift and shift” (later renamed as “move and improve”) and by doing that customers will start to build new applications on top of Oracle Cloud Infrastructure (hereafter OCI).
Even more, Oracle made the case that they’re continuing to learn from all the lessons learned by other cloud providers over the last 10 years allowing them to build something better and faster. It helps that they have a lot of employees with AWS, Microsoft, and Google experience.
Starting with the Oracle Cloud Infrastructure section, Oracle believes that High Scale Cloud items are table stakes – elasticity, pay for what you use, and all the other cloud items you’re familiar with.
Note: you can still run Ravello on top of AWS as well as OCI – it will be less expensive to run Ravello on top of OCI however.
Oracle’s Enterprise Focus is what they believe sets them apart and keeping very tight feedback loops there by listening very closely to customers (backing information cited there). So what does Enterprise Focus mean?
To comment on the slide a bit…
- Meet You Where You’re At = cloud migration without requiring replatforming. There was repeated debate in the room if this was a good thing or bad thing.
- Uncompromised Performance – network virtualization is done outside of the hypervisor by a custom device on the network. This allows use of bare metal and any hypervisor or OS.
- SLAs – more on that later but a big impact.
- Simple, Intuitive Pricing – repeated discussion about avoiding “price shocks” and how they maek that possible. Also proof points around how OCI is less than AWS.
- Completed Integrated Application stack – Oracle was very clear
Architecture consists of an underlying Region with multiple Availability Domains and then goes up the stack.
Some specific notes to call out are….
- No compromises on networking – non-blocking with no oversubscription which makes possible performance based SLA’s. No point in having great compute, storage, or other resources if network can’t keep up.
- Using SR-IOV but can’t disclose some of the tech used for “bump in the network” item that does network virtualization.
- Off-the-box virtualization – network is virtualized outside the compute and hypervisor. Allows the ability to have more flexibility in hardware provided. Some agreement in the room that network virtualization in the hypervisor doesn’t scale.
- Early work on “compartments” to map to organizational structure and billing – there needs to be more details here for enterprise customers.
- Containers aren’t really covered yet – hat tip to Lauren Malhoit. Some discussion but it’s mostly in the TBD category and mentioned in the diagram below.
- Scaling and Performance – lots of discussion in the room around scaling and performance.
There was an interesting summary by Iben that this looks like a cloud built by a database company focused on performance and availability as opposed to a cloud built by a retailer or a search company or an OS company.
For some public futures, Oracle discussed…
…as well as gave some proof points on the speed of cloud capability buildout.
Looking at the roadmap another way, the focus areas are around…
Roadmap focus areas are around…
- Enterprise Lift & Shift – enable enterprise workload migration to cloud without re-architecting. 4 Areas here are…
- Oracle on Oracle – were upfront that most OCI conversations start around Oracle applications.
- Non-Oracle Workloads
- Foundational Solutions
- Governance and Security
- Performance-Intensive – enable workloads that demand large scale/high-performance
- Cloud Native Workloads – enabling new cloud native workloads
To close out this post, I found the stories around “what resonates with customers” to be the most interesting.
- Egress Costs – for one customer 50% of their cloud cost came from egress cost – Oracle’s egress costs is much less as well as offer optionts to totally avoid egress costs.
- Full Environment Snapshots – customers really wanted traditional storage type capabilities to take “coordinated snapshots” to take a full clone of a multi-server environment.
- Hardware Flexibility – the network virtualization allows running the same range of hardware as run on-premises whether small VMs or bare metal servers up to Exadata. 400K IOPs from attached Block Storage with line-rate network throughput makes peope smile.
- SLA’s for Performance, Manageability, and Availability – no one else offers Performance and Manageability SLA’s.
I do want to emphasize the SLA item as found this very interesting.
To go beyond the summary slide, it looks like…
Oracle is showing a continuing commitment to their cloud strategy.
While Oracle has been somewhat quiet about their cloud engineering efforts, they now have thousands of engineers on staff and are also making multiple quiet acquisitions per year.
Although what’s above focuses on OCI, Ravello has also improvied incredibly – I’ll have to write another blog post about that later or leave that to my #RBD2 peers to cover.
I’ll leave you with 2 customer stories.
- YellowDog – Video Rendering as a Service. 10x faster than other options. More details here.
- <Anonymous Customer> (real name but not public) – found Exadata in OCI to be 10x faster than AWS RDS.