Citrix presentation server 4 vs Xen App 5 - comparison documentation
Hi all,
im looking for some documentation (preferably from an independant org) on the differences/advantages of Xen App5 when compared to presentation server 4.
Additionally, some performance documentation on Xen App 5. ie. A server of spec "x" will be able to handle 100 concurrent users for a usage pattern involving the MS office apps only etc etc...
im looking for some documentation (preferably from an independant org) on the differences/advantages of Xen App5 when compared to presentation server 4.
Additionally, some performance documentation on Xen App 5. ie. A server of spec "x" will be able to handle 100 concurrent users for a usage pattern involving the MS office apps only etc etc...
ASKER
i realise i can test, however i still would have thought there would be examples around (with all the appropriate legal disclaimers) - but its still good for a ballpark.
My reasoning for this (and for not wanting to build up my own test environment) is that a client has asked us about their citrix environment (as a part of an overview of their entire network) - We are not citrix people, however they are getting 15 users per server and terrible performance on PS4. While i feel confident in saying "get resonable spec boxes, upgrade to a supported version" and you'll have a better experience with citrix - i find it hard to believe there aren't some example specs/docs around by citrix (which i havent been able to find - hence the post here)
Anyhoo - can you provide a link for the dell or HP tests? or are these not publically available ?
My reasoning for this (and for not wanting to build up my own test environment) is that a client has asked us about their citrix environment (as a part of an overview of their entire network) - We are not citrix people, however they are getting 15 users per server and terrible performance on PS4. While i feel confident in saying "get resonable spec boxes, upgrade to a supported version" and you'll have a better experience with citrix - i find it hard to believe there aren't some example specs/docs around by citrix (which i havent been able to find - hence the post here)
Anyhoo - can you provide a link for the dell or HP tests? or are these not publically available ?
ASKER
also carl, based on your points - im guessing that your a consultant that works with citrix at least a little bit...
So when you go into a new client suggesting to move from PS4 to XenApp5, surely the number of users per box is part of the ROI position you take?
So when you go into a new client suggesting to move from PS4 to XenApp5, surely the number of users per box is part of the ROI position you take?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
yep, ok, fair enough....
not aguing that running it up for yourself gives the truest representation.... but i still think it's odd not to supply some sample scenario's...
I'll definately be suggesting 2008 to this client, so your comment there is welcome.
At the same time you also mentioned - "Now if the client is on 4.0 there are a number of technical improvements included in 4.5 that you would want to push including the more users per server numbers" - you must get that information or opinion from somewhere... i understand people that work in citrix constantly would experience this - but i still find it hard to believe citrix would not publish case studies or ROI reports on upgrades from 4.0 to 4.5 (or 5), of which part would be the number of users per server based on the same workload as used in 4.
But yes, i get it... both of you are basically saying "no, there isnt any such documents" - so either hire an experienced citrix guy or do the testing yourself!
not aguing that running it up for yourself gives the truest representation.... but i still think it's odd not to supply some sample scenario's...
I'll definately be suggesting 2008 to this client, so your comment there is welcome.
At the same time you also mentioned - "Now if the client is on 4.0 there are a number of technical improvements included in 4.5 that you would want to push including the more users per server numbers" - you must get that information or opinion from somewhere... i understand people that work in citrix constantly would experience this - but i still find it hard to believe citrix would not publish case studies or ROI reports on upgrades from 4.0 to 4.5 (or 5), of which part would be the number of users per server based on the same workload as used in 4.
But yes, i get it... both of you are basically saying "no, there isnt any such documents" - so either hire an experienced citrix guy or do the testing yourself!
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
I found the links for HP once and the Dell was in a magazine they used to send out.
Google on HP Citrix performance tests
Most of the links are for older versions of the software.
Google on HP Citrix performance tests
Most of the links are for older versions of the software.
Please keep in mind that the number of 232 users is done on a 64-bit platform with 64-bit XenApp, and 64-Bit Applications. Most companies cannot go to 64-bit for various reasons. You can run Win2k8 32-bit if you still want to go to the 2k8 platform, but do not think that you are going to get more people on a win2k3 or win2k8 32-bit OS.
The problem here is the memory limitations on the 32-bit OS. That the kernel is unable to address more than 2gb of ram, Even with e PAE switch your just not going to get huge numbers. Now this is what you can do very easily:
Get a new server with 16gb or ram or so, Dual quad proc's... the norm., Install XenServer as the hypervisor. Build 4 instances of XenApp on the hypervisor. You now have ALOT more concurrency. For example. If a 32 bit server is getting lets say 30 users on it before we virtualize. XenServer has a 7-10% overhead on the hypervisor when running XenApp on top of it. So for each single virtual instance on the same hardware platform we will get 25 users or so, Then we run 4 or more virtual machines of XenApp on Xenserver and now we are getting 100 users on the same hardware platform. so 30 users up to 100 users on same hardware and its free to do this.
If you can go 64 bit totally, do it and skip the XenServer part. If you need to retain 32 bit, user XenApp on XenServer.
The problem here is the memory limitations on the 32-bit OS. That the kernel is unable to address more than 2gb of ram, Even with e PAE switch your just not going to get huge numbers. Now this is what you can do very easily:
Get a new server with 16gb or ram or so, Dual quad proc's... the norm., Install XenServer as the hypervisor. Build 4 instances of XenApp on the hypervisor. You now have ALOT more concurrency. For example. If a 32 bit server is getting lets say 30 users on it before we virtualize. XenServer has a 7-10% overhead on the hypervisor when running XenApp on top of it. So for each single virtual instance on the same hardware platform we will get 25 users or so, Then we run 4 or more virtual machines of XenApp on Xenserver and now we are getting 100 users on the same hardware platform. so 30 users up to 100 users on same hardware and its free to do this.
If you can go 64 bit totally, do it and skip the XenServer part. If you need to retain 32 bit, user XenApp on XenServer.
No, read my reply. It said Server 2003 x86 Enterprise. That is the 32-bit bit version not 64-bit.
Thats extremely surprising. The users must have been accessing notepad for that test. No way is anyone going to get that number on 32-bit using anything more than notepad.
32-bit Enterprise with 32GB RAM using Office 2003.
ASKER
"Most companies cannot go to 64-bit for various reasons" - i'd love to hear some of these reasons!
Havent run into a place which isnt majority x64 in their data-centre for years.... sure there may be a rogue app here or there, but considering the number of products which are x64 only (and growing) - i just dont see any of those reasons.
Anyhoo - going to close the question - thanks for the replies - not what i was after, but, so be it....
Havent run into a place which isnt majority x64 in their data-centre for years.... sure there may be a rogue app here or there, but considering the number of products which are x64 only (and growing) - i just dont see any of those reasons.
Anyhoo - going to close the question - thanks for the replies - not what i was after, but, so be it....
There area LOT of legacy apps and print drivers used in XenApp environments that just will not work in x64.
Also, running a 32-bit app on 64-bit OS in compatability mode is generally not supported by software manufacturers, and furthermore, Ive seen alot of strange errors in that config.
Carl,
Do you have a link to that white paper? I am interested and would like to read it.
Do you have a link to that white paper? I am interested and would like to read it.
Which white paper?
Check this out: http://support.citrix.com/article/CTX114844
I like these lines:
Similar testing was conducted using this scenario on a 32-bit operating system. The results showed that the system was unable to scale beyond 200 users. This architecture limitation with 32-bit was worked around by using 32-bit and 64-bit applications on a 64-bit operating system.
Note: When sizing Presentation Servers the number of actual users per server varies based on the applications deployed.
I like these lines:
Similar testing was conducted using this scenario on a 32-bit operating system. The results showed that the system was unable to scale beyond 200 users. This architecture limitation with 32-bit was worked around by using 32-bit and 64-bit applications on a 64-bit operating system.
Note: When sizing Presentation Servers the number of actual users per server varies based on the applications deployed.
FInally found the article and Dell sent me a copy of it as a PDF. I have tried my best to create a Word version.
Citrix-Dell-Server-Tests.doc
Citrix-Dell-Server-Tests.doc
XenApp 5 is current.
There are two versions of XenApp 5. One for Server 2003 and one for Server 2008.
Dell and HP have both run user count tests before with Citrix. They found a quad quad core server running Server 2003 x86 Enterprise with 32GB of RAM would max out at around 232 users. Most people can't afford they type of servers they tested.
The standard answer for any question related to Citrix hardware, software or protocols is "it depends". The only way to know for sure is for YOU to test it with your software and test users. You can get an evaluation copy of XenApp 5 for 2003 & 2008 to test and play with yourself.