SplatMan_DK Ars Tribunus Angusticlavius
18y
8,304
Subscriptor++
Confy said:
Nutanix has Veeam support and HPE has Zerto support. I wonder what hypervisor Tesco chose.
Lot's of talk about this in the MSP chats around Europe (and I am a small MSP).
A very good bet is is KVM underneath a private-cloud stack that can handle 40K+ instances in 100+ locations in a fully automated fashion, using Infrastructure-as-code, and which includes baked-in options to run Kubernetes (the Tanzu subscription says so).
Who fits the bill:
Virtuozzo (very good VMWare replacement with K8S but too much proprietary stuff mixed in)
OCI (Oracle are also duchebags so unlikely)
Openstack vanilla (possible but unlikely given the focus on risk - managers want paid support)
RHSOSP (Red Hat OpenStack Platform - expensive but less than half the price of VMWare)
Canonical Charmed OpenStack (cheap and solid option - decent bet especially with Sunbeam for small local sites)
Apache Cloudstack (unlikely as few commercial options would provide support for "at a sprint" migration)
FishOS by Sardina (nice vanilla OpenStack with migration tool, support contract and remote ops but unlikely candidate due to small organizational size for Tescos taste)
Platform9 (cloud-based control plane for on-prem OpenStack and optimized for retail)
Theoretically fits the bill but with lots of re-work:
Rancher (K8S only - unlikely)
SUSE Harvester (Rancher but with KubeVirt VM capability - still unlikely but pretty cool and solid for a long term strategy)
Doesn't fit the bill:
Nutanix (supported by govtools; Veeam - but has good K8S)
HyperV (supported by govtools; Veeam - has no K8S)
Hyperscalers with cloud-only strategy; like Azure, AWS, GCP, Oracle Cloud (too expensive)
Proxmox (great for small shops - but finnicky and absolutely nowhere near Enterprise grade)
Other likely options (the MBA/C-suite approach):
Azure Hybrid environment with Azure Local for most workloads (on-prem) but some Hyperscaler mixed in; for example Azure. Migrating VMWare workloads is easy; doesn't save cost but keeps the few percentages of un-migratable legacy workloads spinning. Not everything in a hybrid environment is easily handled by Veeam and Zerto. But the bulk of VM instances would be covered. AKS on Azure Local for the K8S workloads for example; as well as various PaaS services.
The last option would often be the most appealing for the CIO, the CISO and the risk manager. It offers the easiest migration path - at least on paper. Price would be high but still significantly lower than inflated Broadcom prices; though the first few years of savings would be eaten by the accelerated transition costs.
Me, I'd go for Vanilla OpenStack and hire enough people to keep it running. The offering from Sardina is nice. It supports the accelerated transition requirement, and keeps the exit strategy clean and simple. It also ticks all the compliance boxes upper management wants because paid support is available.
Most vendor lock-in is with Microsoft or Nutanix. Least lock-in and easiest exit strategy is with Vanilla OpenStack, SUSE Harvester, Sardina FishOS, Canonical OpenStack.
Then again ... the suits in 40K+ employee companies rarely make the right tech-decisions. So who knows.
It will spill eventually. :)
