NetApp HCI: More converged than hyperconverged? I have come to the conclusion that NetApp's new HCI system is more converged than hyperconverged: in essence the boxes just add compute nodes and networking to SolidFire (storage) nodes. HCI or CI? Let's row back a little and use IDC's definition of HCI as the starting point for our thinking. Hyper-converged infrastructure ( …
The compute layer can be scaled as much as vCenter can scale, with all the cpu and ram available to VMs.
The compute hardware doesn't run the storage service. The storage hardware doesn't run the compute service.
There is a plugin to vCenter.
Why, then, sell the compute layer? Might as well just continue to let Cisco sell the compute layer.
The plug in in vCenter server needs to know something really special about the SF compute nodes that it couldn't easily do for some other model of compute hardware?
Why wouldn't this plugin be available for the FlexPod SF? Other than 'Cisco compute and nexus switches', what is the difference between the CI based FlexPod SF and the CI based NetApp HCI?
Is Ontap select more HCI than NetApp HCI?
* Who needs such kind of HCI?
4 nodes of select could be 400TB.
I would say, most people?
Could you not put 2 of these 4 node solutions together and get 800TB? Or 3...
* You can use SolidFire vSphere plugin with any servers and it is available with FlexPod SF
What is HCI about NetApp HCI then that isn't FlexPod SF or Select then? Confusing me thinks. Just the ordering method?
Yeah, NetApp is probably facing the most uphill climb on any of the traditional vendors. They have all these new HCI vendors to deal with, including some of their old partners like Cisco and VMware. They still have EMC and the traditional storage vendors out there... on top of all that AWS, Google and MSFT are all fighting to see who can sell object storage for less.
Go read your coworker's (Trevor) article about HCI. People are always the problem. If you create a solution that is going to simplify the data center then who cares how its packaged?
If I was a customer I wouldn't give 2 sh*ts if its truly HCI or not? Can it help me with the problems I have? Yes? Great. If not, then it will fail.
I have been looking at HCI environments for awhile now. really, all we care about is Simple and easy. I mean, really as an admin what I look for in a new platform is for it to make my life easier and save me time. I am still evaluating vsan and I already looked at Nutanix. Neither can accomplish all the tasks we need in our datacenter. If NetApps strange looking offering can be as granular as CI with the simplicity of other HCI then I cant wait to see it. when does it GA?
I agree. A lot of old HCI designs don't really scale and have serious security flaws- for instance having management and data paths on the same plane. I guess with this design, if NetApp can address the simplicity point and get granular scaling then who cares - it's not like HCI is some sort of fundamentalist religion. At the end of the day all vendors are talking about stock CPUs, RAM and disk/Solid State in a metal box - it's the software that counts.
Looks like some one did not look up the definition of HCI.
Hyper-convergence (hyperconvergence) is a type of infrastructure system with a software-centric architecture that tightly integrates compute, storage, networking and virtualization resources and other technologies from scratch in a commodity hardware box supported by a single vendor.
So the question is dose NetApp provide support for VMware if yes this is a HCI. Support from a single vendor is the only check box they might be missing ...
Way to fail...
But don't Nutanix storage only nodes run exactly the same software-defined storage controllers as the nodes that provide compute as well as storage? And aren't all nodes therefore managed in exactly the same way? In my view the presence of software-defined storage controllers on each and every physical node in the cluster (be that in VM-form or a function of the local HV kernel) is a central concept to HCI. With NetApp HCI it seems there are no software-defined storage controllers running on the ESX (compute) nodes, rather it's physical compute nodes talking to physical storage controllers across a network (for all reads and writes) - except the IO travels sideways if you like rather than up/down! If NetApp HCI simplifies data centre operations and scales wonderfully then that's great and those guys should advertise those attributes until the cows come home... but it's still not HCI and NetApp's marketing is therefore misleading / deceiving. Which is really a bit naughty.
Thank you and congrats El Reg/Chris for finally doing your homework to sniff out this marketing BS - you just should have been done in the original article. This is a prime example of some marketing schmuck taking 25yr old technology + all the fancy buzzwords like simplicity, scalability, virtualized, etc... + a fricking vCenter plugin and calling it hyperconverged. Not only that, it's actually 3rd generation HCI! What a joke!! Do they really expect people to buy this shit?? You can't make this up...
Your comment is an example of you not doing your homework. What 25 years old technology you are writing about? Maybe it's better to spend some time on doing some research before writing your comment? HCI is based on SolidFire and it's clustered all-flash storage solution built from ground up.
Granted this isn't Ontap but his point still stands that this is merely a repackaging exercise wrapped in marketecture by Netapp, I'm not really sure there's much wrong with the concept but it's not HCI.
The splitting of storage and compute nodes potentially adds some flexibility but it also removes much of the simplicity, reduces the density and is either going to push up the price for the extra tin required or need deeper discounting and lower the margins vs traditional HCI.
In which case it's much closer to Converged infrastructure, only in this case Netapp's none Ontap version of CI wrapped up to look like HCI.
So NetApp makes ... .just normal old fashioned unconverged garbage, tells you it is converged and gets you to write an article saying that "what we always had before is more converged than since we converged it"? Did you even read what you were writing as you wrong it? How is this different than every pre-converged architecture out there? Pure and Cisco tried to pull this unprofessional marketing crap at the Pure conference a few weeks ago and it didn't fly there either.
This is one of the most ridiculous statements I've ever seen. I'm unclear if this is just technical idiocy or pay to play marketing, but in either case this is a huge problem. If you didn't recognize that you were calling unconverged "converged" then why were you writing about it?
This is just a scale out IPOD. WHo cares? Not even worth an article let alone claiming that the entire industry just changed.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
