Changing the user experience can be dangerous


As many know, I like to do home improvement projects, so I spend quite some time at the local hardware store purchasing supplies.  I’ve been going to the same store for years. I’ve been using the self-service checkout lanes for years. I’ve been purchasing my items with a credit card for years.

Everything was going great until I personally crashed the credit card payment system for my checkout lane (seriously).

So what changed? The credit card checkout process changed because my credit card now includes the security chip technology.  The instructions on the screen were clear, I was simply on auto-pilot, doing the same process I’ve been doing for decades.

Now, don’t get me wrong, I’m perfectly fine changing my behavior IF (and that is a big IF) there is a valid reason for it.  This process changed because it is supposed to make my credit card more secure.

But changing something just to change it with no added value is a great way to annoy me.

Take a look at the following, recently recorded video discussing Microsoft Skype for Business.  Typically, in these types of videos, you see how much better something is to something else.  This is different.  The goal wasn’t to change the experience, it was to make the experience IDENTICAL to what users are accustomed to experiencing.

So tell me, did we hit the mark?

We accomplished this by going back to one of our XenApp Best Practices: For the best combination of user experience and resource consumption, optimize appropriately

So, how does this work?

Let’s look at it from 2 different perspectives:

  1. Admin Perspective: Skype executes within the virtual desktop but the media executes on the end point.  When I make a call to another Skype user, the voice and video goes from my end point to the other user’s end point.  If we are having a conference call, it’s the same process except with more people.  Although Skype is physically running remotely, within my virtual desktop, the media transcoding occurs on the endpoint. This removes network hops. It removes network latency.  It provides a better experience for the user.
  2. User Perspective: The entire Skype session executes within the virtual desktop. But even though the implementation is different, the user’s experience is identical to the traditional PC.

And finally, I keep talking about Skype in the virtual desktop, but that’s because saying published app, shared desktop, pooled desktop, personal desktop, and on and on and on is too long.  This works for all of these delivery options.

Daniel (Follow @djfeller)
XenApp Best Practices
XenApp Videos

XenDesktop 7.7: Windows 7 vs Windows 10


Everyone loves a good battle

  1. The Simpsons vs Family Guy
  2. Kirk vs Picard
  3. Star Wars vs Star Trek
  4. Microsoft vs Apple

And now, we have one more, which will be a big debate for 2016… Windows 7 vs Windows 10

Over the last couple of weeks, I’ve been sharing some of our single server scalability testing results for Windows 7 and Windows 10 across Microsoft Hyper-V and Citrix XenServer utilizing XenDesktop 7.7. Each test compared a different image management model, Citrix policy and operating system optimization configuration.

What was more interesting, and worrisome, was when we started comparing the same test across the different desktop operating systems. The results went something like the following:

  • MCS Winner: Windows 10
  • PVS Disk Cache Winner: Windows 10
  • PVS RAM Cache Winner: Windows 10
  • User Experience Winner: Windows 10

STOP EVERYTHING! This can’t be right. The ProjectVRC.Team did performance tests comparing Windows 7 to Windows 10 and their tests showed Windows 7 always gave higher density. Did we screw up?

Seriously, we were extremely worried that all of the testing was flawed. But with only a few more tests left, we pressed on to see if the results kept favoring Windows 10.

Surprisingly, the very next test changed everything. Windows 7 gave us higher density.

  • Optimized for Scale Winner: Windows 7
  • Optimized for WAN Winner: Windows 7
  • Optimized OS Winner: Windows 7

compare

How can we explain our mixed results vs other 3rd party tests that show Windows 7 winning the battle of server density all the time? As it turns out, most other tests simply focus on cramming the most number of users onto the server while we wanted to also look at User Experience. And Windows 10 outperforming Windows 7 is due in part because Aero (Aero Glass) in Windows 7 is consuming a lot of resources. Windows 10’s version of Aero is more optimized.

When our tests switched from user experience to scale, using the Citrix policy template, Windows 7 Aero was turned off, giving us a huge jump in density. To prove this point, we ran the same Windows 7 test comparing Aero enabled and disabled and the results aligned. But with Windows 10, we didn’t see the same drastic jump when moving between User Experience and Scale, because Aero in Windows 10 is different than Windows 7. For instance, there currently isn’t Aero Glass, which was shown to consume a lot of resources.

So, what will you decide… Windows 7 or Windows 10? Let the battle begin.

And for what it’s worth, my answers: The Simpsons, Sisko (DS9), Star Trek, Microsoft, Windows 10

Daniel (Follow @djfeller)

XenApp Best Practices

XenApp Videos

XenDesktop 7.7 and Windows 10


The other day, I was able to share the latest single server density results when running Windows 7 on XenDesktop 7.7. We looked at a range of parameters like:

      • PVS vs MCS
      • PVS Disk Cache vs RAM Cache
      • Citrix Policies: Very High Definition vs High Server Scalability vs Optimized for WAN
      • Windows 7 optimizations

Once that testing was complete, we moved onto the next version… Windows 10. An again, looking at the exact same parameters.

First, we look at Microsoft Hyper-V 2012R2

HVwin10

Second, we look at Citrix XenServer 6.5 SP1

xswin10

What do you notice?

Between XenServer & Hyper-V… Not much

Between MCS and PVS… Not much as the 5-6% gains in MCS would be offset by increased storage costs due to lack of RAM Caching capabilities

Between the different policies… Around a 7-8% improvement

Between OS optimizations… Around a 7% improvement

The last part I find very interesting.

If you recall, I recently posted a blog (Windows 10 Optimization: The Results) showing that the Optimized OS config, based on the numerous Windows 10 optimizations, showed a 20% improvement in density. As we look at this expanded series of tests, what I now see is something rather interesting. Simply utilizing the Citrix policy templates, we achieve 1/2 of that density gain.

And I can tell you from experience that implementing the Citrix policies are much easier than working through all of those Windows 10 optimizations

So my advice, definitely use the Citrix policy templates as your starting point. If you want to know more about them, I suggest the following:

Server Specs:

      • Processors: Dual Intel Xeon E5-2697 (12 cores each)
      • RAM: 384 GB
      • Workload: Knowledge Worker
      • VM Specs: 2vCPU, 4 GB RAM

 

Daniel (Follow @djfeller)

XenApp Best Practices

XenApp Videos

 

 

 

 

 

 

XenDesktop 7.7 and Windows 7


We recently completed a massive round of testing looking at many of the different deployment configurations we can do with a Windows 7 desktop in XenDesktop 7.7. We wanted to look at how different factors might impact single server scalability.

  • PVS vs MCS
  • PVS Disk Cache vs RAM Cache
  • Citrix Policies: Very High Definition vs High Server Scalability vs Optimized for WAN
  • Windows 7 optimizations

Each test was conducted utilizing the same, knowledge worker workload.

Win7

 

xs

As you see, each test builds upon the previous test while only modifying a single parameter. As I’ve gone through the initial results, some things quickly popped out at me:

  1. Machine Creation Services, from a purely single server scalability perspective, shows some impressive numbers.
  2. Enabling the RAM Cache feature within Provisioning Services gave us a 9% gain in server density.
  3. Every hypervisor tested showed similar percent changes between the different tests.
  4. There was little difference between High Server Scalability and Optimized for WAN because the items within the policy were not a significant part of the tested workload.
  5. Switching from the Citrix Very High Definition User Experience Policy to the Citrix High Server Scalability Policy improved server density by a whopping 30%.

Why did we see such a gain when moving to the High Server Scalability policy? Because the Very High Definition User Experience policy utilizes the H.264 codec, which gives the user a great experience, but at the cost of CPU utilization. When we switch to the High Server Scalability policy, we utilize the latest ThinWire technology. We are able to reduce CPU utilization while increasing server density by changing the way the system encodes the data.

In addition, these policies also modify how items are compressed, how many frames per second are transmitted, plus many more.

To learn more about the Citrix policies, check out the following blogs:

  1. Why you should care about the new HDX ThinWire
  2. HDX Policy Templates

Stay tuned for Windows 10 on XenDesktop 7.7

Server Specs:

  • Processors: Dual Intel Xeon E5-2697 (12 cores each)
  • Workload: Knowledge Worker
  • VM Specs: 2vCPU, 4 GB RAM

Daniel ()
XenApp & XenDesktop Best Practices
XenApp & XenDesktop Videos

Windows 10 Optimization: The untold story


For a few weeks, people in my house kept bugging me to fix our Internet problem. From time-to-time, devices were not able to connect. No eMail. No NetFlix. No Minecraft (this last one was devastating).

I tracked the issue down to my wireless router. The 4-5 year old device was extremely hot. There were 50+ devices connected to it!

After purchasing a new one, all of the connectivity issues went away. As I started looking at the features of the new router, I saw that there is a traffic monitor built-in. So, I turned it on.

The thing I find interesting with this graph at 3 AM.


There is a spike in the middle of the night. I tracked this over a few days and the spike consistently shows up. After doing some digging, it turns out that these spikes are my Windows 10 PCs and my Windows 7 Media Center PC doing their nightly updates.

This brings about an interesting point with regards to Windows 10 optimization. Although we said in Windows 10 Optimization Results blog that the optimizations gave us about a 20% boost in server density, what we have to remember is that many of the optimizations we implemented won’t be accounted for in the test time period.

Take the scheduled task “Customer Experience Improvement Program \ Consolidator” as an example. It runs every 6 hours starting at midnight. In order for this optimization to be reflected in our test results, the test must be run at one of the respective intervals. If my test only runs for 1 hour, there is a good probability I will not have a test running when this task is executing.

Many of the other scheduled tasks run at startup. Most performance tests I’ve seen only focus on the steady state, which means most of the startup scheduled tasks are also missed as part of the test.

So does this mean our 20% benefit for running the optimizations are false? Of course not, but it does indicate that over the course of a workday or workweek, the benefit might be larger than 20%.

But in the end, nothing will ever be better than a real-world comparison.

Daniel ()
XenApp & XenDesktop Best Practices
XenApp & XenDesktop Videos

Follow

Get every new post delivered to your Inbox.

Join 4,289 other followers