The hounds of storage track converged and hyperconverged beasts Tech market researcher IDC's Worldwide Quarterly Converged Systems Tracker has found that worldwide converged systems market revenue increased 10.8 per cent year over year to $2.99bn during the third quarter of 2017 (3Q17) – and that the market inhaled a massive 1.96 EB of new storage capacity during the quarter, up 30 per cent …
This post has been deleted by its author
I live in a hyper-converged environment. I call it a Home.
It has multiple containers, some dedicated to a single function, while others are dynamically configurable. I call them Rooms.
Many of the rooms can work with Windows, but this functionality can be disabled with Curtains.
A few of the rooms directly interconnect, but most use a central network I call a Corridor.
This network is itself multi-level, the connections being made via Stairs. I considered an alternative connection type - a Lift, but found there would be little or no improvement in speed, but requiring considerable restructuring and expense.
Oh, almost forgot, the home has two highly secure connections to the wider environment.
Good summary but seems to be missing is a bit of context and translation.
In integrated systems it seems as if Dell/EMC and Cisco/NetApp are in the lead. However, that Dell/EMC leadership position is VCE with VxBlocks which are really Cisco UCS and Nexus with EMC storage. FlexPods are the same thing (sans the factory prebuild) with NetApp storage. Would it have hurt to state that Cisco has a lock on the high end & integrated market?
As for the hyperconverged market, Dell is obviously reporting VSAN deployments (outside of VxRail) and warping the numbers.
Otherwise, good article.
VSAN alone makes no HCI system as it’s by definition SDS (software defined storage). Granted, since it’s all vmware you could count V’s big orchestration stack (vRealize), Operations Manager, NSX, and of course all the goodies in Enterprise Plus. But how many low and mid level VSAN customers have that?
There is also no hardware management for firmware and settings. With all that in mind I’d call VSAN deployments not hyperconverged, but that’s just my take on things
Umm... what do you mean high end?
Have you seen the performance numbers on Hyperflex, FlexPod, VxRail. Etc...???
My goodness, SQL query times to be ashamed of. MongoDB performance to make a grown man cry. Hadoop performance which looks like someone is taking downers. Object storage numbers of a pathetic nature.
These are low end systems for companies who attempt to compensate for unskilled staff by throwing millions at Dell, Cisco and HP.
I’ll give you a good means of knowing your IT department is incapable of doing anything useful. They actually buy storage systems instead of database systems.
Another clue, they think in terms of VMs and containers. This is a pretty good sign they don’t know what they’re doing.
If you have 10Gbe or faster networking to the servers, you probably have no clue what you’re doing.
If you have servers dual homed to network switches, your system probably is designed to fail and outage windows are scheduled all the time for no apparent reason.
No... these are the low end systems for low performance throw brute force. Unless you are performing oil discovery, mapping genomes, etc... they are about as low end as you can get. Of course, hyperconverged storage is scarily slow compared to specialized storage.
Look at scale out database solutions. They cost far less, require far less hardware and perform far better than what you’re used to. And no... you don’t need VM storage except for your legacy crap which you shouldn’t deploy more of anyway.
That said, VDI is a solution for super big servers... but even then, you shouldn’t have high storage requirements. The base VM should be replicated to every server in the pool and all user storage should be centralized (OneDrive for Business for example). And for that, a simple Windows Server 2016 Core install with Enterprise license and Kubernets should handle it. Though project Honolulu may automate it as well.
Again, no need for storage subsystems, SANs or anything stupid like that. It’s all about the databases.
This makes no sense whatsoever.
First, you are conflating Vblock&FlexPod with VxRail/Nutanix/HyperFlex.
Second, high-end is not necessarily high-performance although I am surprised that you wouldn't consider a big fat Vblock with 100 blades and eight XtremIO bricks or a VMAX AF as high performance.
High-end is enterprise where consistency of deployment, firmware & platform management, and integrated support are far more important than an exotic system that eeks out a little bit of performance here and there. Huge corporations run tens if not hundreds of thousands of VMs. Hundreds or thousands of blades. Yes, the few core systems here and there running databases are treated special but et's be honest, most big shops run Oracle or DB2 on bare metal.
My entire point was that there is something significant missing from this "analysis". Cisco UCS practically dominates the enterprise compute market via converged systems. Look at the market share, it's somewhere around 70%.
So we have Cisco with UCS plus EMC or NetApp owning the enterprise and large commercial market. At the same time we have all these software defined storage vendors coming from the bottom up and de facto introducing a similar management/support concept to the SMB market that enterprise customer enjoyed via converged systems, but with crappy performance and severe limitations.
+++ The question that isn't being asked is, will Cisco use its dominance in enterprise to expand downmarket or will the new players be able to enter the enterprise market? My money is on Cisco but we shall see what happens. +++
Another thing that is not being mentioned at all is the fact that Nutanix has been convincing customers to deploy Nutanix on UCS. Most likely because customers do not wish to move away from UCS. Can't blame them, any sales rep who would suggest ditching UCS for SuperMicro or Dell would be kicked out of our building instantly. That is a significant risk for Nutanix. What's to stop Cisco of suggesting to take three UCS nodes, deploy HXDP and Starship, and take it for a test ride side by side Nutanix? At a significantly more attractive price of course? I think this early Nutanix strategy is going to backfire spectacularly.
Spoken by a true developer who has no idea about infrastructure until the CFO kicks you the whole lot of you out. Your software defined cloud world is nothing else but offshoring of infrastructure and platform at a considerable price uplift. If you were old enough you'd realize that we went through all this in the nineties with the failed concept of Application Service Providers.
".....a bit of context and translation...."
Converged = "Pffft, that's so last year's stuff, it's old and we need to sell newer stuff."
Hyper-converged = "More converged than the old stuff, totally-software-defined, more so than last year's stuff, just take our word for it and buy some!"
Integrated platforms = "What, are you living in the Stone Age!?!?!"
IIRC, we can blame Forresters Research for the moniker "hyper-converged". Still, it is interesting to see Dell leading the charge after all those years propping up the x64 sales charts. Props to Mikey Dell.
And move through the angles of computational space, not the curves. The only way to keep them out of your server room is not to have anything with right angles on it, otherwise they will appear. What! Your room is nothing but racks and cases and enclosures of right angles? Can you not see how the seeds of this madness were sown long ago, before the internet, before computers, before time?
As the definition of a "hyper-converged system" seems to be "completely software defined compute, networking and storage", surely the cloud monsters like AWS are actually the largest HCSs in existence? As far as the cloud customers are concerned, they interface completely through a software interface that portions out compute, software and networking, so surely the perfect HCS (maybe even a super-duper HCS as they do everything remotely through a web interface as well!). I hear Google alone spends well over $2bn per quarter on their cloud, no idea how Amazon spends, but it suggests "DIY" is actually the biggest HCS revenue sector.
No, AzureStack is not included in this. Neither is Google's GCP/GCK Kubernetes stack (currently exclusively with Cisco HyperFlex).
The distinction is as follows:
Local stacks (traditional, converged, hyper converged) need some sort of workload mobility framework to move VMs/applications between clouds. Layer 2 extensions and or IP translation schemas are required as well as some sort of normalization of VM parameters. You either have some powerful software doing this for you (RightScale or CliQr) or you severely limit the cloud targets: Vmware ESX and AWS only, Nutanix AHV and Google GCP only, etc.
AzureStack is completely different. Think of it as a private region of Azure only available to you. You don't manage the local compute stack at all, it's managed out of Azure. You don't control the local OS, it's all part of Azure. In return you get the entire Azure services stack available. This is really meant for developers who can't be bothered to build their own infrastructure.
By the way, from what I hear it's a major pain in the ass to deploy. Everyone thought it would be just a simple click through install and voila! you have Azure on premises. Turns out it takes DAYS to get it to work.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
