Lance Boley:  Good day everyone, we are just about ready to get started ... just waiting on a few more to join.

  Dell-Yinglong:  Hello, Yinglong here from DVS

  Lance Boley:  Welcome ...

  Lance Boley:  Okay lets get started ...

  Lance Boley:  Today we are going to have some experts on to talk about Dell's Desktop Virtualization and what-if's around Office 2013 and Boot on demand.

  Lance Boley:  John, maybe you could introduce yourself and tell us about the team today.

  Dell-John Kelly:  Thanks Lance

  Dell-John Kelly:  I work in the Cloud Client Computing Engineering organization

  Dell-John Kelly:  we are responsible for delivering aoo of our end-to-end desktop virtualization solutions

  Dell-John Kelly:  My main remit is around performance analaysis and characterization (PAAC)

  Dell-John Kelly:  Recently we've looked at 2 interesting topics

  Dell-John Kelly:  Number 1 - what are the density implications for moving to Office 2013 (in terms of number of  virtual desktops per compute host)?

  Dell-John Kelly:  Number 2 - What are the performance implications of booting virtual desktops on demand i.e. not having all desktops on a VDI compute host pre-booted?

  Dell-John Kelly:  Not surprisingly I will concentrate on number 1 first :  -)

  Dell-John Kelly:  We saw some interesting results

  Dell-John Kelly:  Summary is that Office 2013 has a big impact on density compared to Office 2007 and Office 2010

  Dell-John Kelly:  Office 2010 and 2007 are practically identical to each other

  Dell-John Kelly:  Let me just bring up the exact numbers here to give you more specifics....

  Dell-John Kelly:  With Office 2007 or 2010, we can get 116 knowledge users

  Dell-John Kelly:  But guess what - for Office 2013, it's 95.

  Dell-John Kelly:  That's a per host number (R720)

  Dell-John Kelly:  That's approximately an 18% hit on density!!

  Dell-John Kelly:  TWhen I say knowledge worker I mean a Login VSI medium user

  Dell-John Kelly:  Which is pretty much the industry standard knowledge worker

  Dell-John Kelly:  But for Login VSI light (task worker) - something really weird happens....

  Dell-John Kelly:  We got a denisty of 95 also for Office 2013

  Dell-John Kelly:  Comarped to 145 on Office 2007 / 2010

  Dell-John Kelly:  ^ Compared

  Dell-John Kelly:  So we re-ran our tests and scratched our heads

  Dell-John Kelly:  But still got the same, our results were sound

  Dell-John Kelly:  It seems to us that the way Login VSI have put together their workloads, both Light and Medium actually have the same amount of Microsoft Word activity

  Dell-John Kelly:  The difference is the extra activity coming from other acitivites (e.g. flash video, Excel, Powerpoint etc)

  Dell-John Kelly:  But Word is driving the "high watermark"

  Dell-John Kelly:  But I must say it still concerns us that on the latest version of Office, Login VSI 3 gave us this result

  Dell-John Kelly:  But it seems like help is at hand as Login VSI have re-constituted their workloads in Login VSI4

  Dell-John Kelly:  We haven't done any testing on Office 2013 / Login VSI4 yet ... but am eager to get around to it soon

  Dell-John Kelly:  But the main takeaway is - migrating to Office 2013 has serious density implications in terms of per server density

  Dell-John Kelly:  anything from 20-30%

  Dell-John Kelly:  So you need to think hard about whether you need the functionality

  Dell-John Kelly:  I want to give my fingers a rest

  Dell-John Kelly:  Any questions or comments?

  Dell-John Kelly:  Going once...

  Lance Boley:  Tom ... nothing from you today?

  Lance Boley:  or Greg?

  hypervfan:  One thing is doing testing with software like that, but have you seen more memory/cpu usage when testing with just a few users?

  Dell-John Kelly:  memory and cpu scales quite linearly is our experience

  Dell-John Kelly:  But of course as the users log in, the desktops are all pre-booted and sitting idle, so that will take up some cycles to start with obviously

  hypervfan:  And Office 2013 goes up faster than 2010?

  Dell-John Kelly:  in terms of cpu usage, it sure does

  hypervfan:  Is all your testing done with software, or is some manul testing done as well?

  Dell-John Kelly:  we normally login all users using login vsi but we also log in one user manually and that manual user executes sample tasks

  Dell-John Kelly:  because a human being is the ultimate judge of end-user experience...

  Dell-John Kelly:  It's important to say that we also use the Liquidware Labs Stratusphere UX tool to monitor end-user experience

  Dell-John Kelly:  that looks at things like login time, application load time etc

  Dell-John Kelly:  so tools-measured end-user experience, human-measured end-user experience and resource utilization monitoring (CPU, RAM, IOPS and network) gives us a good holistic perspective onthe environment

  Dell-John Kelly:  any other questions? Have I answered all of the questions to date comprehensively?

  Rasmus Haslund:  Hi all

  Rasmus Haslund:  Sorry for the late arrival

  hypervfan:  How is IO on Office 2013 compared to 2010?

  hypervfan:  IOPS

  Dell-John Kelly:  We saw it was a little higher

  Lance Boley:  Hey Rasmus ... glad you could join us

  hypervfan:  Hi Rasmus

  Rasmus Haslund:  What caused the 18% drop in density? Memory usage? CPU usage? IOPS?

  Dell-John Kelly:  The drop in density was due to increased CPU

  Dell-John Kelly:  The IOPS increase was of the order of 10-20%

  Rasmus Haslund:  Was the increased CPU usage across *all* Office applications?

  Dell-John Kelly:  IT was mainly Word

  Dell-John Kelly:  From what we could see, small delta in the others - but mainly Word

  Rasmus Haslund:  What about memory usage?

  Dell-John Kelly:  We concentrated on CPU because that is the gating factor that drove density; I don't have the memory information to hand

  Dell-John Kelly:  If I had to make an educated guess, I would say probably a little higher

  Rasmus Haslund:  For me thats interesting you say that, I have made VDI deployments using Citrix VDI-in-a-box and usually CPU was not 1st or 2nd constraint

  Dell-John Kelly:  Our experience has usually been that CPU is gating factor in our testing

  Dell-John Kelly:  But that is with our standard R720 compute host config of 192GB

  Dell-John Kelly:  So no overcommitment for taks or knowledge workers

  Dell-John Kelly:  ^task

  Dell-John Kelly:  Of course if we went down to 96GB that would probably be different

  Rasmus Haslund:  Did you see if Office 2013 was better at using SMP than Office 2010/2007?

  GC:  are there any plans to test this in XenApp or VMware ThinApp scenarios.   Knowing they are not quite the same kind of products but curious.

  Dell-John Kelly:  We didn't drill down into advanced CPU features such as SMP; but I do know that MSFT are making their Office apps more multi-core /multi-proc aware, so I would guess Office 2013 may make better use of SMP

  Rasmus Haslund:  I dont have more questions on subject #1 for now, looking forward to hearing about #2 :  )

  Dell-John Kelly:  Currently no plans but (not sure if you're internal or external to Dell) - if you feed the request in through the appropriate channels, it may be possible

  hypervfan:  Looking forward to topic 2 as well :  )

  Rasmus Haslund:  We could ask Dennis and Lance to feed it ;)

  Dell-John Kelly:  One other comment - Citrix did similar testing on XenDesktop and got similar results

  Rasmus Haslund:  Did you test this on both Windows 7 and 8? Similar results?

  Dell-John Kelly:  Windows 7 only

  Dell-John Kelly:  One other comment is that it was a one vcpu virtual desktop - industry is moving to 2 vcpus for higher-end virtual desktops and Dell will of course lead that move

  Dell-John Kelly:  Including similar testing to this

  Rasmus Haslund:  I would be really interested to know the results of this test, but adjusted to 2 vCPU

  Rasmus Haslund:  Any chance for a follow-up report in the Rockstar forum on this?

  Dell-John Kelly:  I can't commit at this time but I will take your request onboard

  Rasmus Haslund:  Great :  ) Time is running fast I see... subject #2 hopefully soon :  )

  Dell-John Kelly:  Login VSI's new workloads in Login VSI4 are targeted at 2 vcpus for knowledge worker - so you will most likely see a lot of the results you are looking for very soon

  Dell-John Kelly:  Agreed

  Dell-John Kelly:  So - boot on demand

  Dell-John Kelly:  Scripting and po wer management technologies allow all desktops to be 100% pre-booted on a "just-in-time" basis

  Dell-John Kelly:  But still some customers want to know how their environment will work in a  boot on demand environment

  Dell-John Kelly:  so we tested our local tier-1 solution (tier1 storage - where the virtual desktops reside)

  Dell-John Kelly:  with a number of different scenarios

  Dell-John Kelly:  100% pre-booted (i.e. fully pre-booted); 20% pre-booted and 5% pre-booted

  Dell-John Kelly:  5% is quite aggressive

  Dell-John Kelly:  And logged all users in over the space of one hour

  Dell-John Kelly:  To  replicate a normal morning login storm

  Dell-John Kelly:  And we were actually very happy with our results

  Dell-John Kelly:  Both from the perspective of lstorage atency and Stratusphere UX end-user experience

  Dell-John Kelly:  ^ storage

  Dell-John Kelly:  latency was very comparable across all 3 scenarios

  Dell-John Kelly:  6 milliseconds for 100%, 7 milliseconds for the other 2

  Dell-John Kelly:  which is well below the 20ms latency that's commonly used in the industry as rule of thumb for acceptable latency

  Dell-John Kelly:  Stratusphere UZ also showed consistently good end-user experience

  Dell-John Kelly:  ^ UX

  Dell-John Kelly:  IOPS per user not surprisingly increased

  Dell-John Kelly:  For 100% it was 10 IOPS

  Dell-John Kelly:  For 20% it was 12 IOPS

  Dell-John Kelly:  and for 5% it was 13 IOPS

  Dell-John Kelly:  As expected because the extra booting is an IOPS intensive activity

  Dell-John Kelly:  Also not surprisingly - we saw a change in the read / write percentages for the IO

  Dell-John Kelly:  22/78 for 100% pre-boot

  Rasmus Haslund:  This might be a stupid question, but I assume this is for non-persistant desktops?

  Dell-John Kelly:  37/63 for 5% pre-boot

  Dell-John Kelly:  First it's not a stupid question, I've never seen a stupid question"

  Dell-John Kelly:  Second, it was non-persistent, that's correct

  Dell-John Kelly:  The change in read-write percentage is also as expected - because booting is a read  IO intensive activity

  Dell-John Kelly:  We have 4 minutes left

  Dell-John Kelly:  So questions please

  hypervfan:  read-write percentage is for that hour of testing?

  Dell-John Kelly:  correct

  Dell-John Kelly:  we totalled them across the testing and got the ratio from that

  Dell-John Kelly:  we log all desktops in over an hour, then run them at steady-state for 15 minutes, then log them off

  Dell-John Kelly:  note that this is local tier 1

  Rasmus Haslund:  Maybe a definition of local tier 1?

  Dell-John Kelly:  tier 1 is local to the R720

  Dell-John Kelly:  i.e. sitting behind a PERC

  Dell-John Kelly:  not on a SAN

  Dell-John Kelly:  RAID 10

  Dell-John Kelly:  one other point is that the increased IOPS per user will obviously have an impact on the n umber of users that will "fit" on a SAN (although we didn't test on SAN)

  Rasmus Haslund:  What impact do you expect features such as vFlash to have on these designs?

  Dell-John Kelly:  any flash type functionality will obviously increase capability in relation to number of desktops that can be booted with good performance in a particular time window

  Lance Boley:  We are at the top of the hour and just about ready to wrap it up for today.

  Lance Boley:  So are there any last questions ?

  GC:  Thanks John.  Great job explaining this body of work!

  hypervfan:  Great info

  Rasmus Haslund:  Great chat tonight

  hypervfan:  Interesting to see the 2 vCPU tests when they are coming :  )

  GC:  02121998

  Rasmus Haslund:  Hoping to see John back again another time!

  Dell-John Kelly:  hope you guys found it useful and I will try to incorporate your suggestions into our future work

  Lance Boley:  Alright, thanks everyone for joining today, talk to you next time.

  Dell-John Kelly:  Thanks Rasmis, I hope to be back soon

  Dell-Yinglong:  thanks

  Dell-John Kelly:  Rasmus even

  Rasmus Haslund:   :  )

  Lance Boley:   Thanks John for being our guest today, it was a great topic.