Thoughts from vExpert Session July14th -> Scaling Converged vs Hyperconverged

First published on July 24, 2015 11:23:00 AM at HDS VMware Community

On July 14th we held a private webinar for the vExpert Community. Thanks to the always helpful Corey Romero, vExpert community manager for hosting it and giving us the opportunity to share some of the coolest converged technology out there.

The session was all about the power of the Unified Compute Platform API to allow orchestration of servers, storage, networking via API, the same way you can do so directly within the vCenter UI.

As I mentioned on the webex how about the following service catalog items in vRA !!

  • One item to allow you to deploy a cluster of multiple hosts with the correct ESXi image, hardware profile and host profile attached. So when the process completes you have a working cluster inside vCenter.
    • This is called a service template and this is possible through a single REST call to the UCP Director APi.
  • A second item to add or expand a storage volume (SAN and vSphere) with a single REST call.
    • This call does all the underlying orchestration.
    • If a host needs to be zoned, UCP Director does that for you.
    • It expands or creates the LUN on our arrays, masks to host groups, zones switches and then rescans, creates a VMFS datastore and refreshes the hosts.
    • All with a single API call !!

So this demonstrates what is clearly market leadership in my opinion. Many of our customers want to be able to easily orchestrate virtual and physical resources.

Now they can use UCP Director with either the UI Inside vCenter or HyperV for Business-as-usual operations, or the API for Cloud orchestrated actions. The beauty is that when we talk about orchestrating physical resources from either inside vCenter or via the API, we mean that they could be non-ESXi systems (windows or linux bare metal) or using Hitachi LPARs on x86 blades. Nobody else can offer this to the market. But when talking about vSphere physical servers we mean creating objects with the correct vSphere attributes (delivered via host profile). So much smarter than any alternative solutions.

Here’s a diagram to illustrate parallel orchestration of virtual resources via vCenter API and UCP Director API via a service catalog. Easy peasy !!!.

vRA+UCP+Architecture

 

If anyone needs to know more please contact me.

Scaling converged versus Hyper-converged

From the session, one of items that came up from one of my vExpert comrades was the difference between hyper-converged and converged systems and how these can be scaled. HDS is one of a small number of vendors proudly working with VMware as an EVO:RAIL Solution Provider.

In the Q & A part of the session, one of the listeners asked a few questions about this slide:

 

Screen+Shot+2015-07-24+at+10.54.13

During the session I made a point, based on the slide above, about one of the considerations for selection of EVO:RAIL or any hyper-converged solution being how you scale when you reach the limit of a node. The vExpert questioned this assertion and mentioned that he found it had always been easy to expand hyper-converged infrastructure even within a node. Fair enough but I’m sure many customers don’t want to have to open up boxes and add memory or an additional CPU.

The point I was trying to make is that if I run out of storage on a hyper-converged system, adding another node means adding more CPU, Memory and also associated software licenses for the additional sockets. Conversely you could run out of CPU or more likely memory and also have to add an appliance. In this case you will get more storage which you may not need, as well as the additional licenses.

There is no question that many converged systems like EVO:RAIL are so simple for customers to manage, and there are many advantages of hyper-converged infrastructure. I have also written extensively about the management overhead and complexity managing (some) converged systems.

I can 100% say this does not apply to Hitachi UCP as we have built a layer of abstraction called UCP Director to take this pain away.

So with UCP Director you can easily add hosts for more CPU or Memory, but not add storage. Same applies to bringing on stream more network ports. You won’t need a CCIE to configure them with UCP Director as we have brought in the concept of “managed ports” and “managed servers” within the platform. That’s how we allow bare metal alongside ESXi, by allowing flexible network configuration, while still getting the benefits of shared health monitoring.

We have found the implication of additional nodes, from a licensing perspective is something to consider. As always it depends on the customer’s requirements and in many cases our UCP Director may not best the best solution which is why EVO:RAIL could be the ideal platform, delivered from two of the industry’s most trusted vendors working hand in hand.

What are your thoughts on this ? Feel free to comment !

One thought on “Thoughts from vExpert Session July14th -> Scaling Converged vs Hyperconverged

  1. Hello Paul,

    Thanks for this beautiful article. Most of my doubts with UCP director are now cleared.

    I have a question.

    Recently we have upgrade our UCP director to version 4.0. since then we have not been able to create LUNs (HUS VM) from VSphere Vcenter. It fails after completing 99% of the LUN creation process.

    Any thoughts on this ?

Leave a Reply

Your email address will not be published.