We changed our name from IT Central Station: Here's why
2020-08-11T08:14:00Z

What are the key differences between converged and hyper-converged solutions?


Hi community,

How does hyper-converged differ from converged? Is one better than the other? 

When would one choose converged, rather than hyper-converged? Are there pros and cons to each type of solution?

ITCS user
Guest
59 Answers

author avatar
Top 5Real User

Hyper-converged is typically an "all in one box/rack" solution. It consists of compute, storage & network resources all tied together physically (and through software). 


Hyper-converged for a pro - is a complete solution. You don't have to architect it. All you have to know is how much "power" you need (what you want to do with it). While with converged infrastructure (which can still be 'software defined') you have to match and configure the components to work together. 


More often then not converged infrastructure is cheaper. You might already have the storage and networking resources, for example. And manufacturers put a premium on packaging the solution together. 

2020-08-12T15:25:44Z
author avatar
Top 5MSP

The key differences are scale, complexity and ease of use/management.  


Converged systems are more complex, assembled from multiple vendors using different management tools which allow systems to scale tremendously (Hundreds of servers and Multiple Peta-Bytes of Storage.  


Hyper-Converged systems are smaller, wrapped in a tight, easy to deploy shell that guides the user through system expansion, up to the nodal limits of ~32 servers.  Hyper-Converged systems rely on storage integrated into rack mount systems.  To assure stability, Hyper-Converged user choices are limited on storage and compute options.  These options have been well vetted and are fully supported by the HCI vendor.

2021-12-14T17:02:19Z
author avatarSteffen Hornung
Top 5LeaderboardReal User

@BobWhitcombe 
https://hyperhci.com/2019/06/2... Says there is no node limit for cluster size.

author avatarBobWhitcombe
Top 5MSP

@Steffen Hornung I am not aware of a major provider (HPE, Cisco, IBM, Nutanix, Dell/VxRail) without a 32 node limit on their HCI implementations.  Love to be shown wrong.  You just sent me an outdated Nutanix list of their limitations.  When they were integrating hardware using Dell Servers they had a limit of 32 nodes per cluster.  If you examine the file handling limits and storage capacities - you will find physical support limits due to network latency, RAM overhead, Management overhead and other signal propagation constraints that limit physical scalability.  You will find that 1200 files and VDISK size limit of 62TB don't go very far these days.

author avatarSteffen Hornung
Top 5LeaderboardReal User

@BobWhitcombe The official nutanix site at https://www.nutanix.com/produc...software-options states there is no hard placed limit. Back in 2016 there is a twitter post stating the delivery of a 100-node-solutions using 2 PB storage. Maybe too old for you but nonetheless it is THERE. Limits will most likely be pushed higher. Josh Odgers of Nutanix concluded the sizing limitation claim with "there may be a limit in practicality" and thats it. Don't put numbers out there which are clearly not true. They tend to get stuck in peoples heads.62TB VDISK limit is referring to ESXi 5.5 (and still for 7.0).Node limit for VMware VSAN seems 64 nodes.Nutanix AHV has no specific limit mentioned.Alas I am no product manager for NTNX. But I see no disadvantage in scalability for HCI solutions to get back on track of this discussion.It is in the HCI DNA to scale-out. I am intrigued, though. What are your use-cases for single-VM VDISK beyond 62TB?

author avatarBobWhitcombe
Top 5MSP

@Steffen Hornung Note that in the above document, Nutanix claims a limit of 16 nodes supported under AWS.  Supported node scalability is determined by what the vendor or the underlying software maintains as limits. For example, as you note above, VMWare does not support a VSAN cluster greater than 64 nodes.  HPE's SImplivity, Dell's VxRail (Based on VMWare) and Cisco's Hyperflex all have 32 node limitations.  
As for Enterprise class applications beyond 62TB I see major Epic installations at large, integrated Health Care Organizations and SAP Hana for Major Retailers with Global Supply Chains.  That said - under the pre-COVID cost curve, creating compartmentalized applications that are regional in nature with DR/BC sites makes a lot of sense.  We just saw Ransomware attacks against cloud based Kronos Payroll Systems, killing them just prior to the Christmas Holidays, throwing customers a headache in their operational models at a time when they already had enough issues to deal with - and did not need to face another major Cyber threat.
From a strategic perspective of application design, it makes sense to compartmentalize.  The notion of running your entire business on a single HCI infrastructure is not a best practice and should not be encouraged for the reasons articulated above.  Lose the King, lose the game.  Best to have a lot of pawns as we deal with the current hybrid cloud and cyber space.

author avatar
User

Hyperconverged is a system cluster of at minimum 3 nodes. The system mirrors datas between nodes and runs virtual machines. 


Converged systems is anything between the classic server and hyperconverged platform. This converged concept was useful in waiting for hyperconverged development and should disappear in a near future.



2020-08-12T12:07:02Z
author avatar
User

converged infrastructure still incorporates hardware, running the technology natively on hardware. On the other hand, hype convergence is fully software-defined and completely integrated

2020-08-12T10:23:31Z
author avatar
Top 5LeaderboardReal User

Oh, you cant get rid of hardware in any way. (damn you Apple for auto-correcting english back to german).


But it is true that hci is a software defined approach which has the advantage of delivering new features without new hardware.


Another thing that destinguishes hyperconerged solutions from converged ones is the scale-out nature: simply add more nodes to the system to support new workloads without losing performance because you add all types at once (compute, storage and networking).

2020-08-12T11:36:11Z
Find out what your peers are saying about Dell EMC, FlexPod, Hewlett Packard Enterprise and others in Converged Infrastructure. Updated: January 2022.
563,208 professionals have used our research since 2012.