Oracle – The new Old HPE?

Oracle announced the shuttering of their Ravello Systems solution. Is the inability to capitalize on the $500M purchase an example of Oracle becoming too big to innovate?

Back in 2016, Oracle purchased Ravello Systems. A small cloud startup from Israel. Today, they announced the end of life of the product. The solution enabled users to take their VMware workloads and run them in AWS and GCP. While much slower than production, it was an exciting take on VM management in the Cloud. Many of us in the industry hoped VMware would purchase Ravello and use the IP as a foundation of multi-cloud VM management. Instead, VMware teamed with AWS to announce VMware Cloud on AWS, which is a solid partnership and solution.

Just received notice that @ravellosystems will be End of Life in May. Looks like I need to learn a new infrastructure for all my Cybersecurity training efforts….again 🤦🏼‍♂️

— Charlie Frick (@charliefrick) July 12, 2019

Missed Opportunity? 

Ravello’s initial product wasn’t going to disrupt VMware’s market. Essentially, the solution virtualized existing virtual machines. If that sounds a bit inefficient, you are correct. My experience with Ravello Systems was pretty painful at times. For example, it would take 20 to 30-mins to boot a test vSphere lab environment. However, once up and running, the performance was acceptable for test and development work.

With such poor performance, you wouldn’t use Ravello Systems for production. You don’t have to look much further than VMware Cloud on AWS to see the potential of what Ravello could become. VMware and AWS’ partnership has depth. As an example, AWS created drivers vSphere drivers for the Elastic Block Storage (EBS) service. In theory, vSphere hosts could mount EBS storage from another EC2 instance. It’s no small effort.

Ravello Systems could never have created a “VMware Cloud.” That would require in-depth code level knowledge of vSphere and owning rights to the underlying IP. VMware’s VM format (VMDK) is open enough that Ravello Systems created a solution based on importing that native format.

Ravello Systems could have built on that base. I don’t believe it beyond the stretch of reason, that a different buyer could have worked a deep relationship with AWS. I’ve written that the magic of Ravello Systems was the base hypervisor and management stack. That management stack if developed would provide an alternative for hybrid-cloud management.

Instead of partnering with the major cloud providers and developing deep integrations such as the ones VMware and AWS have developed, Oracle chose a different path. Oracle used the Ravello Systems IP to create a high-performance version of the Ravello Systems solution to run on Oracle’s Cloud infrastructure. I used the service, and my anecdotal testing showed that the new offering more than acceptable for particular workloads. Basically, the native Oracle Cloud compared to a PoC for something like VMware and AWS created.

If a different company than Oracle had purchased Ravello and partnered with AWS, we could have seen a much different take on managing corporate VMs in AWS.

Oracle the new HPE?

Oracle has many of the individual pieces to plumb an alternative to VMC. However, it’s stuck on building Oracle Cloud into an AWS competitor. In addition to Ravello Systems, Oracle has made many next-gen cloudy acquisitions. The list includes

• SD-WAN (Telari) 

• SDN (Xsigo)

• Virtualization (VirtualBox)

• Managed DNS (DYN)

I could write a blog post on each of these areas and the missed opportunity for Oracle or the individual companies to disrupt the hybrid-cloud market. However, Oracle is starting to look like HPE pre-divestiture of business units HPE couldn’t stitch together.

Oracle hasn’t made a strong argument for extending beyond its core competence of enterprise business software. 

Note: I haven’t been able to confirm with Oracle that they will shutdown the service. However, it has been reported that Oracle laid off most of the Ravello team.