Last May, I visited the Oracle Campus in Redwood Shores, to get some deep dive into Ravello’s way cool technology. You can read my initial thoughts here. I was impressed with the solution, and how seamlessly they’d migrated a cloud agnostic mode onto the Oracle Cloud. It seemed to me that the year of quiet since acquisition had proven itself to be a very productive time for the good folks at Ravello.
I wondered what would come of this engineering tour de force in the future, and had a strong desire to see where things would progress. Of course I did! Curiosity is a big motivator for me. Well, this year, over the first week in March, the TechReckoning team invited me back to Northern California to delve into the product for some updates.
In the initial introduction, I was truly taken by the potential for a test/dev environment which could so very easily be built based on the blueprint model. For example, take a clustered app, click and select them from your VMware infra, and create your blueprint. Then take that blueprint and deploy onto the Oracle cloud. Boom, that’s it. You’ll be standing your whole environment up in Ravello’s instance on OCI (Oracle Cloud Infrastructure). Then test your patch, revert to the original blueprint, etc. No need for a sandbox infrastructure any longer. Quick, agile, and simple to stand-up.
But, I honestly didn’t see it as a true prod workload environment. To me, the intrinsic values of VMware have to do with some of the early game-changers: vMotion, DRS, etc. More discrete parts of the VMW architecture like vRealize Ops, Automation, etc. are being considered as part and parcel.
Things have changed. Many tasks they set out to accomplish have been achieved. Let’s say many but not all… To me the fact that they were so up-front and honest about roadmap and goals was so very telling about what they hope to bring to the table, and in addition, the realities of where they are, what they’ve said about their direction toward full adoption tells me much. These folks have worked really hard to accomplish a lot prior to today, I’ve no doubt that they will do so moving forward.
The idea of Lift and Shift (ok ok… Move and Improve, in the Ravello parlance), is valuable in so many ways. Remember, HVX is not a VMware hypervisor, after all. It is a different hypervisor, and as such doesn’t require the VMware licensing costs. It is an Oracle product, though. As such, you’ll not escape a licensing conundrum with Oracle or, for that matter, Microsoft. These are details still being developed. I hope, knowing the quagmire that licensing for Oracle and Microsoft (Sharepoint and SQL server being different than O365, Azure being different than other Microsoft licensing are just a couple examples), that these folks and their legal teams can hammer out an agreement that makes both logical and fiscal sense.
Let’s not limit ourselves here. We saw a great program from Informatica in which they’d built a full suite of online training programs they’d been hosting in their VMW infrastructure, and took that into Ravello’s infra. Now their environment is far more agile, and less expensive for the organization. And they are incredibly happy. We know that more application workloads are being architecture from Informatica are now being loaded to Ravello, and the footprint within and as a part of their data center operations are being expanded. This growth alone tells me that they believe in not only where that product is today, but where it’s going.
Some of the things that I’d like to see are better implementations of backup solutions. While snapshotting has actually a very strong implementation, I’m personally not convinced that snapshots make for a good plan for backups. I would like to get a better handle on the seeding of data into the architecture for Ravello’s infrastructure, how does a database sync up after initial seeding has been established? Note: Not all disc formats or manufacturers support RMAN and/or RSync. I’d like to get a better understanding, or even be involved in that conversation. I’d like to see some expansion as well, maybe some reference designs for how to go about this?
I still see that Ravello and for that matter Oracle as a whole have and have been taking advantage of a great opportunity. Oracle has admittedly entered into the cloud world a bit late. It’s true that this could be seen as putting them in a back-seat role, but in my mind, it’s actually better. They can achieve so much, and yet choose to avoid some of the mistakes made by their predecessors.
I see them doing this already. They’ve built availability zones, and taken into account Data Sovereignty issues, they’ll be building many new ones as well over the coming months and years. They’d not fall prey to the early adopter’s possibility of site to site failover in case of full-site failure.
With the thought process placed into these ideas, and the raw energy these folks have been putting toward hard answers to hard questions, I feel confident that they’ll achieve really good answers.
My gratitude for the Ravello team, The TechReckoning team (@JTroyer, @DailyKat, and @MzAmyWhite), and my fellow bloggers, numerous Illuminati in the space, and with whom I am proud to be associated.
One thought on “Ravello Blogger’s Day 2 (#RBD2) and awesome progress”