| |

Citrix MCS on Nutanix AHV: Unleashing the power of clones!

Update November 2026: This post is from 2016. The core answer still holds: shadow clones do not apply to Citrix MCS on AHV, and they are not needed. I corrected the disk layout and the read path description below. For the current explanation of how AHV keeps MCS base image reads local (per-VM clones and the Unified Cache), read Citrix MCSIO on Nutanix AHV: A Solution to a Problem That Doesn’t Exist.

Citrix MCS on Nutanix AHVWith the release of the plugin for Citrix MCS on Nutanix AHV I got a lot of questions on the inner workings of the AHV cloning so here goes: a blog post that (hopefully) helps you getting a grasp on how Nutanix AHV uses clones in Citrix MCS.

 

As explained in previous articles Nutanix solves data locality issues while using vSphere and/or Hyper-V by implementing Shadow Clones:

The question that I got: ‘Will Shadow Clones help with Citrix MCS on Nutanix AHV too?’

The short answer to this: ‘No’. The longer answer to this question is that shadow clones is basically built to achieve data locality for multireader scenarios like linked clone technologies such as Citrix MCS and the (now retired) VMware View Composer which is not the case for Nutanix AHV.

How does this Shadow Clones thing work again? Here’s an image of one of the slides of the slide deck I presented with Martijn Bosschaart at Citrix Synergy (SYN219):

MCS in legacy infrastructure scenarios:

Key take aways on this image:

  • All reads will traverse the network, congesting your network.
  • Creating a hotspot of reads on your centralised storage.

MCS on Nutanix with vSphere or Hyper-V:

Key take aways on this image:

  • Reads stay local and the primary copy of every write lands on the local node, which improves performance for the user.
  • The reads are no longer traversing the network which opens up more bandwidth for other workloads.
  • Avoiding a hotspot of reads on your centralised storage.

I can almost hear you think “So why won’t Shadow Clones help with AHV?” First of all, it’s because of the cloning mechanism in AHV: AHV’s cloning is super-fast as it does a VM metadata clone and a disk snapshot. Secondly it’s because AHV is using soft pointers to the master disk, this master disk can be accessed locally to the node because of our implementation of cache locality/extent locality and metadata.

That being said, let’s take a quick look at the disk layout on Nutanix AHV:

MCS-based VM disk layout on AHV:

Key takeaways on this image:

  • On AHV, each MCS VM gets its own thin, copy-on-write clone of the base image vDisk, plus the MCS identity disk. There is no shared differencing disk chain for shadow clones to act on.
  • It pre-populates some of the assigned storage.

The first time a VM on a node reads a block of the base image, that read may go to the node holding the data. After that, the Unified Cache on the local CVM serves the block, so every catalog VM on that node reads the base image locally. Michael Webster has real world data reduction numbers for this kind of workload in this post.

Concluding:

This solution provides data locality out of the box when it comes down to the cloning mechanisms used by Citrix MCS on Nutanix AHV. If we combine this with the powers of the storage-native actions that we have available on Nutanix AHV with Acropolis and Prism we get a Citrix XenDesktop (now Citrix DaaS Local) on Nutanix solution that provides a single, high-density platform for desktop and application delivery. This modular, pod-based approach also enables deployments to scale simply and efficiently with zero downtime.

Get new posts by email

One email when a new post goes up. Nothing else.

I only use your address to send new posts, and every email has an unsubscribe link. See the privacy policy.

ShareLinkedInXEmail
Kees Baggerman

Kees Baggerman

Senior Technical Director at Nutanix · NVIDIA Enterprise Platform Advisor · Former Citrix CTP · 15+ years in EUC

Kees Baggerman is Senior Technical Director, Performance & Solutions Engineering R&D at Nutanix, where he leads a global team responsible for defining how enterprise applications are delivered on the Nutanix platform. An NVIDIA Enterprise Platform Advisor since 2017 and a former Citrix Technology Professional, he has spent 15+ years driving EUC strategy and technical direction across architecture, product, and customer success. He has been writing here since 2011, sharing what he learns at the intersection of platform engineering and enterprise IT.

2 Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.