Hyper-converged prospects? Thanks to firm-gobbling habit and OEM deal, Dell has 3 Dell has three horses in the hyper-converged system race: VMware’s VSAN colt, EMC’s ScaleIO mustang and OEM’d Nutanix’s filly. Come 2017, VMware and EMC should be inside Dell and eyes will look quizzically at the Dell Nutanix OEM deal. Specifically, Dell says it has the broadest hyper-converged stable offering available from …
I'm guessing one can make a case to include the CPS-P as a hybrid solution. So make it QTY 4, perhaps QTY 5 w/ Dell HCS for MSFT - Standard.
Overlaps in corporate capabilities are good until customers get confused and start running full speed in the other direction, which are API's and Cloud Service contracts.
Just as EMC and Nutanix are in battle for HCIA dominance, Dell gets further under the covers with Nutanix. Curious to see if other products in the Dell/EMC arena will support Nutanix (Acropolis) like Avamar. There is definitely a push from Nutanix for customers to "rip and switch" from vmware.
If you are not considering hyper-converged technology today, you will likely be left behind tomorrow. With SSD capacity growing at an awesome pace, $/GB dropping at the same time, and technologies like 3D XPoint coming to a datacenter near you in the next year, why would you NOT buy into hyper-converged? External arrays are for cheap-and-deep cold storage in two years, connected via 10/40/100Gb iSCSI/NAS. Anything needing performance sits in a server-SAN with SSD's. Not suprised to see this rapid growth and product development.
Atkinson says Dell is conservative about who it will partner with while a couple of para above they list out 5 different HCIA products. There is so much redundancy across the Dell - EMC acquisition that after all the redundancies (including employee's) are eliminated so will shareholder value.
What shareholder value? Dell is a private company.
What is missing on all Hyper-converged stacks today are E7 based 2 and 4 socket systems for those larger footprint apps (SAP/Oracle/DB2).
It will also take some time before Hyper-converged catches up with data services when compared to a standard SAN approach, but they will get there.
BUT as with all things IT we will come full circle. Once the hyper-converged stack becomes "Standard" a new SAN like technology will emerge or fashions will change and we'll start re-distributing Compute/Network/Storage again.
So re E7 CPU – are these workloads really a workload of the future?
With our dev friends pushing containerisation as apps get re architected in a micro services world the scale up requirements of a monolithic app world start to diminish.
This is where the scale out approach of HCI with API access to a software defined infrastructure makes so much sense.
I actually agree that some 4 socket options would be handy for customers that like the approach but have legacy apps that still require the CPU/memory footprint of an E7 approach – they would like to adopt HCI, but are struggling to refactor their apps.
What “data services” do you think are missing from todays HCI solutions?
And can’t help but chuckle and think you’re correct re full circle.
#iwork4dell
Dell selling VXRail could be a strange one as it is currently not on Dell server hardware - EMC/VCE I believe use Quanta hardware -this would see Dell effectively reselling another hardware vendor - sure it wouldn't be too hard to switch it out with a similar Dell server through.
From what we see, the level of data services that a number of the leading SDS and HCI vendors have are sufficient for the majority of virtual use cases. Only real gaps are in applications which require bare metal, have a dependancy on Fibre channel, the application vendor won't support the use of de-duplication/compression or the application require synchronous replication to a secondary site.
Other than the above they seem to be more than fit for purpose.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
