As IT leaders move away from VMware, they face a critical decision: do they stick with traditional storage architectures, or is now the time to finally unlock the full potential of an infrastructure that converges virtualization, storage, and networking technologies?
Early convergence efforts centered on hyperconverged infrastructure (HCI), where storage ran as a virtual machine under the hypervisor, commonly called a vSAN. While adoption has lagged behind traditional three-tier architectures, recent advancements have significantly improved vSAN, making it worth reconsidering by addressing past shortcomings.
The above is an ad, but it both open about it and technically interesting, if a bit repetitive.
I honestly had no idea a vSAN was primarily running on hidden storage VMs on the host. I thought the storage stack was as tightly integrated as networking/compute at the hypervisor level, so the light breakdown there was worth the read.
Not all of them. Ceph on Proxmox and (iirc) VMware vSAN run bare metal. That statement was a call-out post for Nutanix, which runs their storage inside a VM cluster. Both of these have been doing so for years.
Okay, that tracks better. Im familiar with ceph and promox using it as a “fake” vSAN. Im also familiar with Vmwares vSAN and had never seen any indicator of an internal storage VM, so that was odd.
It being Nutanix doing the above makes sense as I haven’t worked with them yet.
Out of curiosity why would you call Ceph a fake HCI? As far as I’ve seen, it behaves akin to any of the other HCI systems I’ve used.
Id call it viable hyper-converged infastructure when it in use as such like with proxmox, but its not scoped to just being a vSAN. Its a distributed storage network. its design is way wider than just being used for HCI/vSAN/etc.
I have a small cluster running using Starwind for my vSAN. For me it’s much cheaper than a hardware equivalent and is performant enough.
Oh and I haven’t had a “stop work” issue with it in 8 years.
Somewhat remarkably it was OK performance-wise when sync/iSCSI traffic were running on 1Gb copper connections to spinning rust storage… Now I have 10Gb fibre between the hosts, coupled with nvme drives, and it’s quite (comparatively) quick.
As with all things YMMV… But vSAN is the way for my use case.
“comparing VMware vSAN, Nutanix AOS Storage and VergeIO VergeOS”
Hint: There is no vSAN/UCI/HCI silver bullet.
They all suffer from performance related issues either up front or over time, have scalability deficits or limitations and lackluster price to performance ratios (Nutanix can suck my 5yr forced upgrade balls).
Did small Nutanix deployments a while ago, and liked it enough to go all in on vSAN for the datacenter about 7 years ago.
No regrets. Great performance, no san storage, no San network. Easy to manage on dell readynodes.
Going to go all in with Nutanix next round. Same costs but it’s not Broadcom.
Ceph / Openshift Data Foundation is also an option for hyper converged clustered storage.
Ceph is also quite tightly integrated on Proxmox for smaller deployments.
Yup! Ceph is quite nice once you’re up at the ~6 node scale. Openshift Data Foundation is based on Rook, which is a nice way to automate Ceph deployment, as well.
I’ve actually been running Rook on my home server rack, up to 120TiB now 😁
