PVS and MCS
PVS runs on dedicated servers (VM or physical) and is pretty heavy on network io and should be designed properly while MCS needs nothing more than hypervisor snapshots which, by know, we already have.
When we are going to implement physical targets for both VDI and RDS we need PVS. MCS integrates on the hypervisor storage layer and therefor can’t be used on physical targets. This is the simplest decision of them all!
One decision is about the expectation of boot or logoff storms. If there are we need Provisioning Services. The reason for this decision is based on IOPS. Provisioning Services is more IOPS effective then Machine Creation Services. But is this still valid now with the release of a XenServer feature called Intellicache? Intellicache caches the vDisk on local hypervisor storage (like cheap SSD) and therefore offloads the IOPS issue.So is this decision stil valid? The answer is yes and no! If you use anything else then XenServer use Provisioning Services. If you use XenServer you can go either way.
Hyper-V Generation 2 vms cant be used with xendesktop. MCS cant be used with standalone hardware devices like blade pcs. MCS generates 1.6x or 60% more IOPS compared to PVS.
mcs equal read/write. pvs reads are more. raid 0 fast read
- Storage foot print would be very small for PVS.
- Using cache on ram, would give a better performance on overall VDI performance.
- Citrix Admins have complete control over user’s VDI.
- Windows updates in normal scenario take lot of network bandwidth and storage – With PVS, only one image should be updated, rest of all are done with a reboot.
- With dynamic VDIs, we can power manage them as per user shift timings.
Pros to Non-Persistent Dynamic VDIs.
- Patches and software installs/updates only installed on base image and pushed out to all machines with reboot. Minimizes time to update and impact on Storage/Network.
- Machine reset to “Factory” on reboot. If a user manages to install a piece of software it is removed on reboot.
- Not as many machines required since users can share machines. Instead of 3000 machines for 3000 users we would more likely need 2000 or maybe even less to service the same 3000 users.
- 2000 VDIs only use 100GB(size of vDisk on PVS server) of total Storage instead of 200TB
- Can be power managed so that only VDIs that are in use plus a few to cover new logins will need to be running instead of all machines running at all times.
Both technologies can provision new VMs.
- MCS is a component within xendesktop and PVS is a separate component in citrix.
- PVS is a rapid VM provision technology. It doesn’t check whether you use those VMs as standalone VMs, or add them in citrix etc.. With MCS you can only use those VDIs within a citrix site.
- When you create a catalog using MCS, it copies the snapshot to all datastores. This is as per MCS architecture, but with PVS you don’t have that problem. Citrix says from version 7.15 and above, this problem is rectified.
- When you create 10 machines with MCS each having 100 Gb drive, considering thick provision, you will spend 1000GB just for spinning up those machines. But with PVS, considering cache on ram, you can have 100 VMs with no hard disks at all. You will only spend 100 GB for your base VM. That’s it.
- This is applicable for both MCS and PVS. with non-persistent machines, you will have complete control on your end VMs. Users cannot install random applications, they are removed on reboot.
Want to learn more on Citrix Automations and solutions???
Subscribe to get our latest content by email.