Earlier this month, I was extremely excited and honored to participate
in my first Oracle Ravello Blogger Day (RBD2) event, which brought together 29
influencers from around the globe along with a set of Oracle customers and a
select group of Oracle employees. The
event was an all-day session which took place at the Oracle HQ in Redwood City,
CA and was put together by John Troyer and the TechReckoning team.
This was my first time at the event, now in its
second year. I was unable to attend the inaugural event 10 months ago, which unfortunately meant I missed out on much of the backstory and
prior discussions. Rather than spending too much time rehashing
what was already communicated to the room in the previous year's session, the Oracle
team basically started where they left off and continued with the new story. While lacking some of the context, I knew enough to
understand what was going on and simply had to play catch-up with some of the
terminology and underlying technology which wasn't difficult thanks to a group
of industry experts sitting at the table with me who had attended the previous
year.
For a one-day session, you might think there wouldn't be a lot of information covered, but you'd be wrong. The day went by fast, but we were drinking
from a firehose as one person after another took to the virtual stage to
enlighten us in the ways of Oracle Cloud and Ravello. Several employees shared the
company's vision, strategies and areas of focus from Ravello to the Oracle
Cloud Infrastructure, backed up by slides, live demos and a lot of Q&A and
open and honest discussion. The day was capped off by proof points from two
actual customers who went into more detail about their own experiences with implementing Oracle's technologies.
For
those not familiar with Ravello or Oracle Cloud, here's a quick rundown.
Ravello
Ravello Systems was acquired by Oracle back in February of
2016. The company was founded in 2011 by
a team that created the KVM hypervisor. The Ravello team took its virtualization
know how and created what it called a "nested virtualization
powered cloud service" which wraps complex application environments into
self-contained capsules in order to run them on any cloud.
Delivered as-a-service, Ravello enabled enterprises to recreate entire
application environments with existing VMware or KVM virtual machines and
complex networking, to be deployed on any cloud, without any changes. As
an example, Ravello announced at AWS Re:Invent 2014 that it could clone an
exact replica of a customer's entire VMware vSphere environment, and then transfer
that environment to Amazon Web Services or Google Cloud Platform. Once
moved, Ravello could spin-up that complete environment in less than 20
minutes.
It's that idea behind grabbing those
on-premises virtualization environments and the construct of easily migrating
or creating Hybrid Cloud environments that caused Oracle to scoop them up and
bring them into the Oracle Public Cloud organization.
At RBD2, it was nice to see that Ravello not only survived
the acquisition but is thriving in its new Oracle world. Too many times, we've seen good and
interesting technology get gobbled up by a much larger entity, only to see it
become "shelf-ware" for one reason or another.
In this case, Ravello continues to be worked on as an
ongoing product line and it has become a valued and key component of the much
larger Oracle Cloud story.
Now that the Ravello team has a native cloud to focus its
development efforts against, the group has worked hard to bring the portability
of its HVX technology to Oracle's Cloud Infrastructure. And the results are impressive.
Oracle Cloud
Playing Catch Up
Let's face it, Oracle is coming a bit late to the party. The big three, Amazon AWS, Microsoft Azure,
and Google Cloud are way ahead; but Oracle, based on what I've seen and heard
at the RBD2 event, is moving at a fast pace to try and catch up.
One area for them to achieve success is to first work with
existing Oracle database customers, a track that makes sense. You dance with the partner that brought you,
and these clients are all too familiar with Oracle as a company and
what it provides. For Oracle, it's
a land and expand type situation, where you offer something additional to an
existing customer base. Oracle helps
move these customers into the 21st century of Cloud in a way that is
familiar to them, keeps them operational and allows them to meet SLAs; and it does
so at a price point that doesn't destroy the budget. It also helps with the old "one throat to
choke" model, which probably isn't very PC to say these days. For existing Oracle clients, this cloud
offering could also be sold as an "easy button" to natively host those complex
systems in the Oracle cloud without having to re-factor things.
While this strategy may be great for moving existing Oracle
clients to the cloud, that alone isn't going to bump Oracle up to the #3 spot
let alone allow them to challenge Amazon or Microsoft for market share
dominance.
Vision and goals
Oracle has set an ambitious goal for itself in building its
second-generation cloud infrastructure: to create an infrastructure that
matches and surpasses the performance, control and governance of enterprise
data centers, while delivering the scale, elasticity, and cost-savings of public
clouds. The result, Oracle Cloud
Infrastructure was built from the ground up to be an Enterprise Cloud, equally
capable of running traditional multi-tiered enterprise applications,
high-performance workloads, and modern serverless and container-based
architectures.
Oracle's vision: "Customers can run their entire business on our cloud."
Time and time again, the theme heard throughout the day was "Oracle
Cloud Infrastructure is built for enterprise."
As part of their proof points to explain that concept, Oracle shared the following
slides:
By the numbers,
Performance Intensive Workloads
Oracle is hanging its hat on the
value it provides on the performance side of things. For organizations that have dipped their toes
in the water of the public cloud and seen inconsistencies in performance across
different layers, this is a major bullet point.
Without geeking out too much, the
Oracle Cloud compute instances are purported to be powered by the latest Intel
Xeon Processors and secured by an advanced network and data center architecture
designed for high availability and enterprise class governance:
- Up to 52 hyper-threaded cores
per instance for ultimate performance
- Regions available in US and
Europe
- Up to 51 TB of local NVMe
SSDs, as well as advanced block, file, and object storage options
- Industry-first
non-oversubscribed network with dual 25 Gbps network interfaces per host
- 400K IOPS per machine and less than 1
millisecond latency
- X7 GPU shape with 2 NVIDIA P100 GPUs
The company has also created an impressive list of feature enhancements
in a rather short period of time, all things considered:
Part of performance has to do with the relative closeness of
the cloud server to the endpoint. To
compete with other public clouds, Oracle is also going to have to expand their
geographical footprint of coverage in order to meet user demand much like Microsoft
and Amazon have been doing. Oracle said
they are working hard at this as well.
Other core themes moving forward include:
-
Enterprise Life & Shift - Enable enterprise
workload migration to cloud without re-architecting
-
Performance-Intensive - Enable workloads that
demand large-scale, high performance
-
Cloud Native Workloads - Programmable infrastructure
for cloud-first development
With Lift & Shift, Oracle wasn't simply talking about seamlessly
migrating Oracle workloads to their cloud, they were also talking about
moving
non-Oracle workloads as well, such as Microsoft apps, SAP, and VMware
(using Ravello). They also talked about working out security
and governance aspects, to be compliant with all the leading standards.
Enterprises need Service
Level Agreements
With its database software business, Oracle is no stranger
to the Enterprise space. And they know, if
you want to play in the Enterprise, you have to have SLAs. And that's extremely true and important in
the Cloud market.
For a large company, if something goes wrong and they lose
access to hardware or software for even a split second, they can lose
money. For something like a hospital
facility, it could mean lives. This is
serious stuff, and Oracle clearly understands that need. Enterprise organizations want SLAs and guarantees
around services hosted in the cloud. And
Oracle is accepting that challenge by providing SLAs for their Storage, Network
and Compute performance and manageability, not just simply uptime.
Peace of mind with an SLA and a guarantee goes a long way to
provide customers with the confidence needed to make a move to the public cloud. It's one of the key problems public clouds
face and Oracle is addressing it head on to capture the attention of enterprise
organizations.
Performance is great,
what about Price?
The idea here is not only to make pricing simpler, but also
cheaper.
Allowing a customer to move their existing on-premises stack
as is to an Oracle Cloud is not only an attractive option from a management and
basic understanding, but it adds an inherently cost saving approach to the move
as well.
Other cost savings come into play with things that aren't
normally thought of like networking egress. Oracle said they are trying to eliminate charges that seem to add up out
of nowhere for a cloud user. One of those being networking egress charges, which
Oracle said they aren't charging for data transfer when using Oracle FastConnect. Oracle FastConnect is a network connectivity
alternative to using the public Internet for connecting your on-premises data
center or network with Oracle Cloud Infrastructure. It should be used when you require higher
bandwidth options that can scale with your business needs, provide 99.95%
availability and offer a more reliable and consistent networking experience
when compared to Internet-based connections.
Stay tuned for more pricing discussions in the coming weeks and months, but Oracle hinted that
it will be a much cheaper alternative to the likes of Amazon and
Microsoft.
Customer Use Case
There were also two customer presentations at the end
of the day, explaining how they use Oracle's technology, its implementation and
the gotcha's they encountered along the way.
Evidently at last year's event, the conversation was around a
test/dev environment, which could easily be built on Ravello's blueprint model. Build your blueprint from a VMware environment,
and then quickly and easily deploy it to an Oracle Cloud Infrastructure. Then, perform your dev/test operation, and
revert back to the original blueprint when done. No need to build an expensive sandbox environment.
Fast forward, and this time the discussion moved on from
dev/test to production and training. Informatica University discussed building
out a training portal. The portal makes
it easy for them to quickly stand up and deliver instructor led and self-paced
training sessions for their clients. With it, they are able to deliver more
than a hundred different self-paced courses and 45 instructor led courses. It
not only sped up deployment, but it also reduced their training costs by 70%.
Conclusion
In all, Oracle is on the right track and clearly focused on
their objective. I believe they know
they have a hard row to hoe, but the progress being made is impressive. There are still plenty of holes that need to be addressed,
but they have a roadmap (under NDA) that they are building toward to try and get them there. They've made great strides in the last 10 months
since RBD1, but all eyes are on them to see if they can keep up that momentum.
By starting with a group of known elements, in this case, an
existing customer base of Oracle software users, the company is forging ahead
by building something that works for those enterprise users and leveraging it for expansion
to other similar enterprise buyers.
If there is an RBD3, I hope to be brought back so that I can
see and hear first hand what type of progress they've made on the things we
weren't allowed to discuss, yet.
Stay tuned.