A total silence… i’m back!

When i looked at my blog https://florenttastet.wordpress.com and realized that it has been over 2 months i haven’t posted anything, all of a sudden i realized that time flies. Well let me correct that.

 The last 2 months have been focussed on two things: training & certifications! 

This journey brought through the most advanced technologies available and a deep dive in it. Quiet an adventure! Today i’ll be going through PVS and MCS. Next week CBRC.

Citrix Provisioning Services

If you are a Citrix customer, you know about it. If you’re not, i hope you’ll learn something.

Despite its internal competition, “Machine Creation Services” or “MCS”, PVS is well established in Citrix strategy to deliver on demand target OS at a fraction of the storage cost. 

 Acquired from Ardence Inc in 2006, it is clear it became an important player in the Citrix portfolio today. Thanks to the “brains” at Citrix, who forecasted the next challenge: storage consumption.

Citrix Provisioning Services (“PVS”) comes in a addresses the most challenging road block of virtual sprawl by centralizing one component and streaming it on demand, allowing an automated provisioning (and re-provisioning) of our servers (and virtual desktops)

 Originally PVS was build by Ardence Inc to provision servers in the datacenter. Interesting, since Citrix aimed at provisioning Datacenter “SaaS” solutions hosted on servers, what a great idea it was to acquire a technology that would allow you to deploy “ghosted” images of OS and installed Application in it.

 And indeed, the primary usage of PVS is about Xenapp; in these days of Software Defined DataCenter, PVS imposes itself as a key element of a true automated Datacenter solution, and its usage can be extend to Desktop Virtualization.

 How does PVS works?

In a nutshell, using Provisioning Services management console, you’ll have to prepare a device (called “master target device”) for imaging and install any required software on that device. 

A resulting vDisk image will be created from the master target device’s hard drive (that you just configured) and saved to the network on a storage device, server or storage array.

Now the magic operates: once the vDisk is available from the network, the target device no longer needs its local hard drive to operate; it boots through PXE, directly across the network. PVS streams the contents of the vDisk to the target device on demand, in real time. The target device behaves as if it is running from its local drive. So the disk comes from the network, and the processing occurs on the target device. 

Cool no?

How this helps me or my customers?

Server provisioning: imagine for a minute that you have a virtualized Xenapp farm with packaged or not applications to be serviced and consumed by your users or clients. You would be able to keep the servers in the farm to a minimum while automatically provision additional Xenapp servers when the demand increases for the applications you’ve published.

Desktop provisioning: making its way in that sphere, PVS, would allow you to centrally managed your gold image and provision Virtual Desktops on demand. Helping overcome the well known “boot storm”, my question to you would be: How many desktops do you wish to provisioning ‘cause that will not be the issue.

Rule of thumb for PVS

Hum…. with MCS in the picture, PVS, while able to scale down, is typically used for larger deployment such as 2500 + deployment (yes i know Citrix says 5000+ but let’s put things in perspective: i am in a SMB market). PVS is focussed on streaming; who says streaming says network; who says network says bandwidth, hence why i like it.


Citrix Machine Creation Services

It looks like to me that Citrix realized that PVS could be in some cases an overkill, mainly because it created the need to have additional components to manage but more, adds a layer of complexity that shades the benefits of PVS and creates a dependency on two major components: Server OS and Streaming service.

Let me explain.

The Windows Server OS could fail; you might say: “yes but  in such case if properly configured the virtual machines would fail over to the other Provisioning Server” and you would be absolutely right, but…

The ultimate concern is the one that affects the Citrix Streaming Service (hosted by the OS). It is that service, responsible for “streaming” the requested image blocks across the network has a bug and crashes on one Provisioning Server and the virtual desktops fails over to the other Provisioning Server, if that same circumstance (that caused the crash) occurs which caused the virtual desktops to crash the streaming service on the first PVS server, then it will probably crash the streaming service on the second PVS server…and now you’re in trouble… 

So… in late 2010, beginning of 2011, Citrix released MCS.

How does MCS works?

First things first: MCS is a embedded component of XenDesktop, so no need to have externals components such as… PVS (oups)

XenDesktop MCS can create two types of desktops:

  • Pooled desktops are the “disposable” stateless desktops that best practices dictate should make up most VDI deployments. User data is not stored on the desktop, the desktop is reset to it’s original state when a user logs out.
  • Dedicated Desktops are similar to a traditional desktop deployment. Each user is assigned a dedicated desktop. In this case all the changes made to a desktop are retained from session to session.

Does that makes you think about someone else in the VDI world?? say what? View? You got it! LOL

Joe Keegan has made a very simple overview of MCS that i particularly like:

It all starts with a VM which you plan to use as your Master Template. All the desktops in a catalog will be based off a Master Template. When creating the catalog you select the Master Template and choose if you want to create a pooled or dedicated desktop catalog.

MCS will then take a Private Snapshot and a public snapshot of the Master Template VM. The public snapshot is visible in the vSphere Client and is called Citrix_XD_CatalogName (Where CatalogName is the name of the catalog). The private snapshot is not visible via the vSphere client. The Public Snapshot can be removed without impacting the catalog and MCS.

Once the Private Snapshot is made then a Clone is created from that Snapshot. Each catalog will create it’s own snapshot and it’s own clone. So if you have two catalogs based on the same Master Template then you will have two clones. Also a clone is created for each data store included in that catalog. So if you have two catalogs each using the same five data stores then you will have ten clones, a clone for each catalog on each data store.

Each Desktop in a catalog is granted read only access to the clone on it’s data store. An Identity and a Difference vDisk are “layered” on top of the read only clone. The Identity vDisk includes all the specific desktop identity information such as host name, password, etc. The Difference vDisk is where all the writes and changes to the desktop are stored.  These Identity and Difference vDisks for each desktop are stored on the same data store as their related clone.

The difference between a Pooled or Dedicated desktop is what happens to the Difference vDisk when a user logs out.  In a dedicated desktop the differencing disk is retained after a user logs out while in a pooled desktop the differencing disk is deleted. 

Where do i play with MCS?

I don’t like to read “PVS vs MCS” as i think they both have a key play in a true automated Datacenter solution, but, indeed they have their pros and cons. PVS is mainly driven by storage concerns, MCS by network concerns. ouch!

One thing to consider about MCS is only Hosted VDI would be available on a (agnostic) hypervisor in the data center ONLY. 

So it means, no Blade PC, so streamed image and no hosted shared desktops.

it that so much of a big deal? It depends. 

My take on this topic is typically MCS would do the job especially when you’re looking at hosted desktop solutions. If there is a need to stream the desktops “away” from the datacenter, that you have or not a large infrastructure, PVS is to be considered and therefore, eliminates the need of MCS. 

You have one “to do”: look out for Xendesktop7. Many answers are provided…

So in a nuthsell, what should you do?

I use the PVS-MCS decision tree. I like it. It drives my decisions and helps me align and focus. I use it as GUIDELINES only, so the final decision is yours, but it truly helps.Image

take a look at it and leave your feedback; i’d like to hear about you and what you live on the day-to-day basis. Knowledge has that strength that it always needs to increase and improve. I thank you in advance for everything you may want to share.

On that note, i wish a wonderful weekend fulfilled with relaxing time, family time, nice weather and lots of ice cream.

i’m now up for my most preferred moments of the weekend:  spending breakfast with my beautiful 21 months princess!

cheers friends.


About florenttastet

As an IT professional and leader, my objective is to help an organization grow its IT department with new and innovative technologies in order to have production at the most efficient level ensuring the right alignment in the deployment of such technologies through a precise Professional Services results in a extraordinary experience for the customer. Team member of multiple projects, I have developed a strict work ethic allowing development of superior communication skills, as well as the ability to multi-task and meet precise deadlines. As an IT veteran with a broad background in consulting, management, strategy, sales and business development, I have developed an deep expertise in virtulization using VMware and Citrix products with a strong skillset on Storage Arrays (HP, EMC, Netapp, Nimble & IBM). I have also developed a Security practice through CheckPoints NGX R65-R66 (CCSA obtained) and Cisco PIX-ASA product line. Specialties: Microsoft infrastructure products; Monitoring HPOV, SCOM, CiscoWorks, Firewalls: checkpoint, PIX and ASA. Virtualization with VMware (ESX through vSphere & View/Horizon), Microsoft (Hyper-V server, VDI and App-V), Citrix (Xenserver, XenDesktop, Xenapp), Storage (EMC, HP, Netapp, Nimble & IBM), Reference Architectures and Converged Datacenters (vSPEX, Flexpod, vBlock, PureFlex & HP Matrix)
This entry was posted in Uncategorized. Bookmark the permalink.

One Response to A total silence… i’m back!

  1. Pingback: A datacenter journey

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s