PVS vs MCS – Part 2: Scalability

This is part of a series comparing Provisioning Services and Machine Creation Services

• Part 1: Resource Delivery Options
• Part 2: Scalability
• Part 3: Storage Optimization
• Part 4: Deployment
• Part 5: On-going Maintenance
• Part 6: Architecture
• Part 7: Summary

In Part 1 of the PVS vs MCS debate, we saw

• Provisioning Services bridges the gap between the physical and virtual world
• Machine Creation Services bridges the gap between the on-premises and cloud world

Let’s continue digging into the PVS vs MCS debate and focus on scalability. Scalability plays a big role in the solution. If one solution only scales to 500 users, it would be a bad idea to select that option for a deployment of 5,000 users.

It seems like more people question the ability for Machine Creation Services to scale than they do Provisioning Services. When Machine Creation Services was initially released, many believed it was only for small deployments of around 500.

So, let’s look at how these two technologies work, which will give us some insight into their scalability potential.

Provisioning Services

Provisioning Services utilizes network streaming. A master image is contained within a single file. Each Provisioning Services server accesses the read-only file, and streams portions of the file to the target devices (FYI booting Windows 10 to the logon screen only requires 250MB of streamed data!).

What will bottleneck? Your Provisioning Services server’s NIC.

If you need more scalability, you can increase the NIC speed, add more NICs, add more servers or do a combination.

Machine Creation Services

Machine Creation Services, on the other hand, is not based on a stand-alone server install like Provisioning Services.  The Machine Creation Services functionality, contained within the XenApp and XenDesktop controller, is based on interacting with the hypervisor storage (local and shared). As virtual machines are created and updated, the Machine Creation Services commands are sent to the virtualization hosts.

So, what will bottleneck first? Your storage.

Each storage cluster can only support so many virtual machines before it cannot handle the incoming requests.  If you need to add more scalability, you need to create additional storage clusters.

Comparing the Two

Both solutions can scale, as long as you add more storage clusters or more servers. But one thing you should keep in mind is that the user experience, or how well the target device performs, is based on different factors:

• Provisioning Services links user experience to the stability and performance of your network
• Machine Creation Services links user experience to the stability and performance of your storage

But in terms of scalability, where does that leave us in the PVS vs MCS debate?  Even

Daniel (Follow on Twitter @djfeller)
XenApp Advanced Concept Guide
XenApp Best Practices
XenApp Videos