The Problem with Repatriation
On paper, repatriating workloads to the data center makes sense. In practice.
Moving to the public cloud is a frictionless infrastructure experience. I don’t need to talk to a salesperson. I don’t even need to configure a VPN between my infrastructure and the public cloud. I can load an appliance onto my VMware vSphere cluster, walk through a wizard, and the automation tool will start migrating my virtual machines.
I shut off the lights to my data center, and I’m “migrated.” Very oversimplified, but the motion is proven.
Fast forward a couple of years. If I stopped there, I’d find that my cloud infrastructure bill is outrageously expensive. Naturally, I’d look to a colocation provider to repatriate some workloads. The process is as simple as… Wait, that’s where we begin to see the problem.
Traditionally, I’d have to contract with a colocation provider, order hardware, rack and stack, and the list goes on. If you are a born-in-the-cloud engineer (BTCE), the experience is an anti-pattern. The friction of going back to the data center is too great.
The folks at Equinix have been trying to get me to try Equinix Metal. I have to say, I was pretty impressed with what’s there so far. I can deploy a data center with dedicated hardware via self-service without talking to a soul. There’s much more to be desired, but the experience shows promise for the world's Dell APEX and HPE GreenLakes.
Specifically, if I could deploy an HPE dHCI platform or DellEMC VxRail using Equinix Metal’s self-service, I think the friction to move back on-prem would be significantly reduced. Now Day-2 management… that’s another discussion.
Share your thoughts by replying or commenting. Would better bare metal self-services increase hybrid cloud adoption?