Up until now, we have had 3 major players in the public cloud: Amazon Web Services (AWS), Microsoft Azure Cloud, and the Google Cloud Platform (GCP) have long been seen as the major players, with, in many cases, VMware playing the infrastructural back end of the virtual servers sitting within. Not exclusive, mind you, as standing up virtual environments in any of these can be done, but for enterprise organizations, with already existing VMware architectures in place, the desire to not reinvent the wheel, but extend their VMware infrastructure to a cloud provider has been a huge motivator to migrating their platforms toward the cloud. I have long heard in my customer conversations about a desire to move to the cloud, a lesser desire to continue building, adding, supporting and maintaining physical infrastructure, I’d rather let the hyper-scalers do it.
Now, we’ve all heard tales of organizations repatriating their workloads off of the major cloud providers and bringing them back into the on-premises environment. The logic almost exclusively has focused on the cost of these functionalities. I believe when functions like data egress charges, processor ratchet, network bandwidth utilization, and the like become chargeables, we start to see how this cost utilization becomes a more viable rationale for repatriation. As always, in my conversations with customers, caveat emptor. Try, as best you are able to keep your eyes open to your options before jumping. You may actually be surprised by what you find out.
Over the last couple months, I’ve been working with the CTOAdvisor, (@CTOAdvisor @MrsCTO) alongside industry stalwarts @DemitasseNZ, @Geekazine, @ThomTalksTech, and @CXI to build out an infrastructure on the newly launched Oracle Cloud Infrastructure, in an effort to evaluate their infrastructure in comparison to that of the other providers in a specific use-case. We established a scenario many organizations have been facing, which feels particularly relevant in the Covid19 world. Essentially, there’s a real need to ramp up certain functionality to accommodate the increased volume of remote workers needing access to internal systems. We decided an existing VMware based datacenter, with a small deployment of Horizon View, to support their occasional remote workers now needs to expand the footprint of the VDI environment to support those users now working from home. To do this, the logical step would be to create a hybrid environment between a cloud provider and the on-premises Horizon View deployment.
With this as our goal, we determined we’d do the appropriate due-diligence and test out all the options available to us, including AWS, Azure, GCP, and the newly launched Oracle Cloud initiative. We wanted to establish how this could be achieved on each of the platforms, how much learning a VMware administrator would require, what kinds of flexibilities would be available to that environment, and how much we could truly tweak the management layer to make things as seamless as we could. Speed to uptime was important, but what we really needed was stability, and as much similarity to how business had been done on-premises, within the CTOADC (CTO Advisor DataCenter). Our ideal scenario would be to vmotion the horizon view environment to the targeted cloud provider, create a stretch cluster, and manage the Horizon View environment through vCenter, as if it all were housed on-premises.
All the processes were videoed by Jeffrey Powers (@Geekazine) with technology being driven by @ThomTalksTech, and the videos hosted by Alastair Cooke (@DemitasseNZ). All this will be posted on the CTO Advisor (https://www.thectoadvisor.com/) and on the brilliant BuildDay Live youtube channel.
My job was to attempt to document the process, the difficulties, and the successes, along with creating the narrative we were establishing around our rationale. This surely was an undertaking, and we enjoyed the process.
Different than other analysts, our goal was to provide evidentiary proof of the process. We do believe that when we approach an analysis, just reading the documentation provided by the vendor is insufficient. The CTO Advisor feels that doing the legwork, and showing that lends credence to the process, and will make a report more valid. This is in no way disrespectful to the other analyst groups out there, but we wanted to build-out a case, and then prove the inherent value.
The next posting will be a summary of the report and highlight the findings.