The concept of cross cloud isn’t new and in truth a lot of vendors today are working to, or have solutions that aim to convert workloads from one platform to another. Zerto do this with their Cloud Fabric with the ability to move certain VMs from ESXi to Hyper-V, AWS and Azure and every combination in between. Veeam also have a new feature where you can restore ESXi or Hyper-V VMs to Azure…again, limited in functionality but a strong indication of what’s to come given the latest Veeam announcements.
Both Zerto and Veeam market their solutions well, however those that have been involved in V2Vs know that only under certain conditions do conversions go smoothly. There is no doubt this cross platform world is getting more reliable and more and more vendors are chasing the perfect conversion. However what Veeam and Zerto are offering is Backup and DR services that complement VM workloads either on-premises or in a cloud…the end game with these products isn’t mobility…its availability.
Focusing back on VMware it was clear to almost everyone that the Cross Cloud Platform featuring Azure and AWS workload migrations, was tech previewed to show that VMware is relevant in an enterprise multi cloud world but I am going to argue that the focus on the VM as the base unit of measurement is misguided…especially when it comes to VMware supporting it’s vCloud Air Network providers. I understand it as a necessity being able to have a class of portable applications in this new microservice and serverless world while having them transportable between multiple clouds. Again, I don’t believe the VM should be the base unit of measurement and the unit shown to be the most transportable.
Service providers need to play to their strengths, which in the vCAN world is no bill shock fixed cost IaaS workloads. This remains the base platform for a significant portion of any on-premises or cloud workload. Service providers take most of their revenue stream from compute, storage and networking that are the building blocks of instance based and resource pool offerings from which VMs can be provisioned and consumed. If you ask any service provider they would say that they would like total VM stickiness and any mechanism that aims to make VMs more portable will impact the bottom line and threatens ongoing viability.
Having customers access a VMware provided console that moves VM workloads off VMware based infrastructure and onto AWS or Azure to my mind is close to madness, and while there is an argument to suggest that cloud is the new hardware and VMware want to manage this new hardware…it still doesn’t make up for the fact that most revenue is made by having VMs staying local and not having an easy way to migrate them to platforms where smaller margins are the norm.
Going back to the point of this post around the theory that the VM shouldn’t be the base unit in a cross cloud world, I believe that for the sake of the vCAN VMware should be focusing within the VM and the applications that run within them…working towards a truly hybrid scenario whereby Platform and Feature as a Service offerings are managed, configured and operated via the Cross Cloud platform. This will help achieve a sustained revenue stream for IaaS providers that in truth, still represents the best value for money for the vast majority of critical business applications that are in existence today, all while allowing consumers the choice of going out and finding the best “As a Service” offering that specifically suits application requirements.
At the end of the day I do wonder which side of the VMware business wins out…the one that derive their revenue from Enterprise…or the one that derive their revenue from Service Providers. Unfortunately I know where the bigger revenue streams lie and that doesn’t bode well for Service Providers. It’s all about the corporate dollar after all.