[Micronet] VDI?

classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

[Micronet] VDI?

Bill Clark
I was discussing Virtual Desktop Integration with some co-workers, and we
got to wondering if IST had a service offering in that area.  All I can
find from searches are references to a meeting held to discuss the pilot
for such an offering "currently under development" as of July, 2009:

http://ist.berkeley.edu/tam/projects/docs-vdi

Did anything ever come of this?  Does IST offer any sort of Windows
Terminal Server (or other) solution?  There is some renewed interest
within our unit, but we'd prefer to use an IST-hosted solution if
possible, rather than go to the trouble of setting up our own server for
testing.

Since the slides from the above mentioned presentation indicate that the
technology was to be revisited "when it is more mature" in 1-2 years,
maybe I should be volunteering to participate in a new pilot?

-Bill Clark
Systems Unit
Graduate Division


 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] VDI?

Michael Sinatra-2
On 4/13/11 1:13 PM, Bill Clark wrote:
> I was discussing Virtual Desktop Integration with some co-workers, and we
> got to wondering if IST had a service offering in that area.  All I can
> find from searches are references to a meeting held to discuss the pilot
> for such an offering "currently under development" as of July, 2009:
>
> http://ist.berkeley.edu/tam/projects/docs-vdi

That's all you could find?  Your google-foo is broken. :)

Since Ryan Means has left UCB (oops! so have I), and he was the chair of
the VDI working group, I think it's best to point you to his report,
which is linked from this article:

http://inews.berkeley.edu/articles/Oct-Nov2009/DVP

It gives you a lot of the technical and implementation issues that were
uncovered.

If you search goofle for 'IST Desktop Virtualization Project', the first
two non-ad links are the article and the report.

michael

 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] VDI?

Gabriel Gonzalez
Hi,
There are a number of us still on campus that participated in that
pilot. If the report raises questions, please send them along.

Thanks,
Gabriel Gonzalez
School of Law

On 4/13/2011 1:21 PM, Michael Sinatra wrote:

> On 4/13/11 1:13 PM, Bill Clark wrote:
>> I was discussing Virtual Desktop Integration with some co-workers, and we
>> got to wondering if IST had a service offering in that area.  All I can
>> find from searches are references to a meeting held to discuss the pilot
>> for such an offering "currently under development" as of July, 2009:
>>
>> http://ist.berkeley.edu/tam/projects/docs-vdi
>
> That's all you could find?  Your google-foo is broken. :)
>
> Since Ryan Means has left UCB (oops! so have I), and he was the chair of
> the VDI working group, I think it's best to point you to his report,
> which is linked from this article:
>
> http://inews.berkeley.edu/articles/Oct-Nov2009/DVP
>
> It gives you a lot of the technical and implementation issues that were
> uncovered.
>
> If you search goofle for 'IST Desktop Virtualization Project', the first
> two non-ad links are the article and the report.
>
> michael
>
>  
> -------------------------------------------------------------------------
> The following was automatically added to this message by the list server:
>
> To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:
>
> http://micronet.berkeley.edu
>
> Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.

 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] VDI?

Jon Forrest
In reply to this post by Bill Clark
I've been thinking about VDI. I don't
think it's all it's cracked up to be.
I just don't see what its advantages
are.

For example, at one time it was thought
that VDI would save money on hardware
because a user VDI "station" would be
less expensive than a true PC. Given
what's happened to the cost of generic
x86 PCs and LCD monitors, I doubt this
is true anymore.

It was also thought that VDI would
make it easier to deploy a large number
of seats because the amount of work
necessary per seat would approach
zero. Given what's happened in
network booting, software deployment,
and roaming logins, I doubt this
is true anymore.

It was also thought that VDI would
reduce software licensing expenses
since it would only be necessary to
license software running on the
server. Since some (many) packages
couldn't tell when they were being
run remotely, it might have also
been possible to avoid other licensing
fees. It's turned out that it isn't
hard for software to detect that
they're being run remotely, and then
refusing to run unless the proper
license exists. For example, Matlab
works this way.

It arguable that the added complexity
that VDI requires on the server side
negates any added simplicity on the client
side. Plus, at the end of the day, if
a user has trouble running Word or Excel
on a standard PC, they're still going to
have the same problems in a VDI environment.

(Many of these issues were mentioned in
the IST report).

What would you be trying to accomplish
by switching to VDI?

Cordially,
--
Jon Forrest
Research Computing Support
College of Chemistry
173 Tan Hall
University of California Berkeley
Berkeley, CA
94720-1460
510-643-1032
[hidden email]

 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] VDI?

Bill Clark

> It arguable that the added complexity
> that VDI requires on the server side
> negates any added simplicity on the client
> side.

True, but that's why I was asking specifically about an IST-hosted
solution.  Server maintenance can be centralized far more cost-effectively
than can client support.  You may be right that it's still not worth it
overall, though... I guess it would depend on just how much more
complicated (and expensive) the server side becomes.

Mostly we were looking at the effort involved in migrating end users to
new hardware, particularly when done on a temporary basis so that the
original hardware can be upgraded/repaired.  Even the best system cloning
tools take more time and effort than a VDI-based approach.

The report was a bit disheartening, so I doubt we'll pursue this much
further at this point... but I expect that eventually the technology will
mature to the point that not just applications but the desktop itself will
run in "the cloud" and clients will be simple and interchangeable.  Just
not now, I guess.

-Bill Clark


 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] VDI?

Julie Ashworth
I will provide a success story, but I'm not sure how
applicable this is to Bill, since we use linux.

I support computing needs for researchers at the Helen
Wills Neuroscience Institute. Their relatively unusual
requirements include large amounts of disk space (to
store images), and graphics-accelerated applications
for image manipulation.  
For years, we used linux on the desktop. The biggest
problems were frequent hardware failures, inefficient
use of resources, relatively slow network (for NFS),
and noisy offices.
We consolidated resources in the data center last
October. We use NoMachine's NX clients and the freenx
server for remote desktop access. A batch processing
system helps balance load. The compute 'cluster' now
consists of 11 high performance workstations with 100
cpu cores and 1TB of RAM. The 150TB of disk is
available via NFS on a private network (up to 2Gbps).

The benefits of remote desktop (and consolidation)
include  more efficient use of resources, less
physical hardware to troubleshoot and maintain, and
increased uptime. Researchers prefer the remote
desktop, because resources are consolidated, and
they can suspend/resume their desktop session from
anywhere that supports ssh. Also, since researchers
can choose the desktop operating system of their
choice, we see fewer systems on the desk (consuming
power and bandwidth).

The sysadmin time increased initially, mostly to
troubleshoot graphics-acceleration problems, but has
dramatically reduced. Note that desktop systems
(mostly windows and Mac) are maintained by the
researchers themselves.
 
The user FAQ is here:
http://www.neuro.berkeley.edu

Best,
Julie


On 13-04-2011 13.50 -0700, Bill Clark wrote:

>
> > It arguable that the added complexity
> > that VDI requires on the server side
> > negates any added simplicity on the client
> > side.
>
> True, but that's why I was asking specifically about an IST-hosted
> solution.  Server maintenance can be centralized far more cost-effectively
> than can client support.  You may be right that it's still not worth it
> overall, though... I guess it would depend on just how much more
> complicated (and expensive) the server side becomes.
>
> Mostly we were looking at the effort involved in migrating end users to
> new hardware, particularly when done on a temporary basis so that the
> original hardware can be upgraded/repaired.  Even the best system cloning
> tools take more time and effort than a VDI-based approach.
>
> The report was a bit disheartening, so I doubt we'll pursue this much
> further at this point... but I expect that eventually the technology will
> mature to the point that not just applications but the desktop itself will
> run in "the cloud" and clients will be simple and interchangeable.  Just
> not now, I guess.
>
> -Bill Clark
>
>
>  
> -------------------------------------------------------------------------
> The following was automatically added to this message by the list server:
>
> To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:
>
> http://micronet.berkeley.edu
>
> Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
---end quoted text---

--
Julie Ashworth <[hidden email]>
Computational Infrastructure for Research Labs, UC Berkeley
http://cirl.berkeley.edu/
PGP Key ID: 0x17F013D2

 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] VDI?

Kevin Newman-3
In reply to this post by Bill Clark
Please note that the report was based on an older product version.

Not saying that the high level issues around VDI aren't still there but that
the specific VMware issues might have been addressed.

Kevin Newman
Associate Director, Infrastructure
Information Technology
University of California, Berkeley
University Relations
2000 Center Street, MC 4200
Berkeley, CA 94720-4200
510.643.5930  FAX:  510.643.2154
E-Mail:  [hidden email]


-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of Bill Clark
Sent: Wednesday, April 13, 2011 1:51 PM
To: Jon Forrest
Cc: [hidden email]
Subject: Re: [Micronet] VDI?


> It arguable that the added complexity
> that VDI requires on the server side
> negates any added simplicity on the client side.

True, but that's why I was asking specifically about an IST-hosted solution.
Server maintenance can be centralized far more cost-effectively than can
client support.  You may be right that it's still not worth it overall,
though... I guess it would depend on just how much more complicated (and
expensive) the server side becomes.

Mostly we were looking at the effort involved in migrating end users to new
hardware, particularly when done on a temporary basis so that the original
hardware can be upgraded/repaired.  Even the best system cloning tools take
more time and effort than a VDI-based approach.

The report was a bit disheartening, so I doubt we'll pursue this much
further at this point... but I expect that eventually the technology will
mature to the point that not just applications but the desktop itself will
run in "the cloud" and clients will be simple and interchangeable.  Just not
now, I guess.

-Bill Clark


 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe
from its mailing list and how to find out about upcoming meetings, please
visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and
the list's archives can be browsed and searched on the Internet.  This means
these messages can be viewed by (among others) your bosses, prospective
employers, and people who have known you in the past.


 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] VDI?

Tom Holub
In reply to this post by Jon Forrest
I think Jon must be posting from the new Commodore machine (40 columns...)

On 4/13/11 1:40 PM, Jon Forrest wrote:

> I've been thinking about VDI. I don't
> think it's all it's cracked up to be.
> I just don't see what its advantages
> are.
>
> For example, at one time it was thought
> that VDI would save money on hardware
> because a user VDI "station" would be
> less expensive than a true PC. Given
> what's happened to the cost of generic
> x86 PCs and LCD monitors, I doubt this
> is true anymore.

It remains true that thin clients are cheaper than true PCs, both in
terms of life cycle hardware cost and life cycle support cost.  They
have fewer failures and longer life spans than PCs.

> It was also thought that VDI would
> make it easier to deploy a large number
> of seats because the amount of work
> necessary per seat would approach
> zero. Given what's happened in
> network booting, software deployment,
> and roaming logins, I doubt this
> is true anymore.

You're right, but for the wrong reason.  The reason VDI has limited
appeal is that the amount of work necessary per seat to manage the OS
and installed software really doesn't look a lot different than managing
separate physical boxes.  Unless your virtual desktop image is totally
locked down and cloned, you still need BigFix or an equivalent, you
still need roaming profiles, etc.  So you go from managing 200
individual clients to managing 200 individual images.  You gain a little
bit by being able to more easily migrate an image from one piece of
hardware to another, but you still need a lot of support for individual
desktop customization.  (Or else, why would you not just run Terminal
Services?)

--
Tom Holub ([hidden email], 510-642-9069)
Director of Computing, College of Letters & Science
101.D Durant Hall
<http://LSCR.berkeley.edu/>


 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] VDI?

Jon Forrest
On 4/13/2011 2:52 PM, Tom Holub wrote:
> I think Jon must be posting from the new Commodore machine (40 columns...)

No, it's just I have a very narrow personality.

> It remains true that thin clients are cheaper than true PCs, both in
> terms of life cycle hardware cost and life cycle support cost.  They
> have fewer failures and longer life spans than PCs.

Can you give a reference to these statistics?
What's your definition of the hardware differences
between a thin client and a PC?

> You're right, but for the wrong reason.  The reason VDI has limited
> appeal is that the amount of work necessary per seat to manage the OS
> and installed software really doesn't look a lot different than managing
> separate physical boxes.  Unless your virtual desktop image is totally
> locked down and cloned, you still need BigFix or an equivalent, you
> still need roaming profiles, etc.

Let's say you use a network boot
to get the OS started, group
policy to load software, and then
roaming profiles to store per-user
configurations.

> So you go from managing 200
> individual clients to managing 200 individual images.

I don't see this. You don't have 200 individual images - you
have one standard boot image that has the latest service pack
and updates. You do have 200 roaming profiles but that's
a simple matter of disk space.

I am presuming that the software you want on the desktops
can be installed by group policy or other similar methods.
I'm also presuming Windows 7 as the only desktop OS and
an adequate network infrastructure.

I agree that older versions of Windows don't have what's
necessary to live in the world I'm envisioning.

> You gain a little
> bit by being able to more easily migrate an image from one piece of
> hardware to another, but you still need a lot of support for individual
> desktop customization.  (Or else, why would you not just run Terminal
> Services?)

Terminal Services might indeed be necessary if you're
forced to run older software.

Jon



 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] VDI?

Tom Holub
On 4/13/11 3:10 PM, Jon Forrest wrote:

> On 4/13/2011 2:52 PM, Tom Holub wrote:
>> I think Jon must be posting from the new Commodore machine (40 columns...)
>
> No, it's just I have a very narrow personality.
>
>> It remains true that thin clients are cheaper than true PCs, both in
>> terms of life cycle hardware cost and life cycle support cost.  They
>> have fewer failures and longer life spans than PCs.
>
> Can you give a reference to these statistics?
> What's your definition of the hardware differences
> between a thin client and a PC?

A modern thin client has no moving parts; flash RAM, convective cooling.
 I don't have stats but it seems obvious that the configuration is more
reliable and longer-lived.

>> You're right, but for the wrong reason.  The reason VDI has limited
>> appeal is that the amount of work necessary per seat to manage the OS
>> and installed software really doesn't look a lot different than managing
>> separate physical boxes.  Unless your virtual desktop image is totally
>> locked down and cloned, you still need BigFix or an equivalent, you
>> still need roaming profiles, etc.
>
> Let's say you use a network boot
> to get the OS started, group
> policy to load software, and then
> roaming profiles to store per-user
> configurations.
>
>> So you go from managing 200
>> individual clients to managing 200 individual images.
>
> I don't see this. You don't have 200 individual images - you
> have one standard boot image that has the latest service pack
> and updates. You do have 200 roaming profiles but that's
> a simple matter of disk space.

It's not just roaming profiles; for most use cases you wind up having
per-image software installation, which means you have per-image software
patching.  For the use cases where you don't have per-image software
installation, it's not clear what VDI gets you over Terminal Services.

>> You gain a little
>> bit by being able to more easily migrate an image from one piece of
>> hardware to another, but you still need a lot of support for individual
>> desktop customization.  (Or else, why would you not just run Terminal
>> Services?)
>
> Terminal Services might indeed be necessary if you're
> forced to run older software.

The question is, what are you gaining by running VDI instead of Terminal
Services?

--
Tom Holub ([hidden email], 510-642-9069)
Director of Computing, College of Letters & Science
101.D Durant Hall
<http://LSCR.berkeley.edu/>


 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] VDI?

Curtis Salinas
In reply to this post by Bill Clark
Hi Bill,

There is no IST-run VDI service, nor is there currently an effort to develop one.  There are benefits to running your desktop in the virtual environment if you find that terminal services do not meet your needs, and we have a few customers who have had successful deployments doing that, mostly on Windows 7.  This does not offer the extra features that come with VMware View or Citrix's XenDesktop, but does offer independence from client side hardware for those that wish to go this route.

For Virtual Private Server (VPS) pricing, "server" being a misnomer in this case, requests for quotes and orders can be placed through our provisioning portal:

http://estimator.berkeley.edu

We maintain a standard Windows 7 image, so you would just select "Window 7 Template" as the template when you get to that point in the order.  If you are interested in trying this out, I generally recommend a second virtual CPU for a multi-purpose desktop.

- Curtis

> -----Original Message-----
> From: [hidden email] [mailto:micronet-list-
> [hidden email]] On Behalf Of Bill Clark
> Sent: Wednesday, April 13, 2011 1:13 PM
> To: [hidden email]
> Subject: [Micronet] VDI?
>
> I was discussing Virtual Desktop Integration with some co-workers, and we
> got to wondering if IST had a service offering in that area.  All I can find from
> searches are references to a meeting held to discuss the pilot for such an
> offering "currently under development" as of July, 2009:
>
> http://ist.berkeley.edu/tam/projects/docs-vdi
>
> Did anything ever come of this?  Does IST offer any sort of Windows Terminal
> Server (or other) solution?  There is some renewed interest within our unit,
> but we'd prefer to use an IST-hosted solution if possible, rather than go to
> the trouble of setting up our own server for testing.
>
> Since the slides from the above mentioned presentation indicate that the
> technology was to be revisited "when it is more mature" in 1-2 years, maybe
> I should be volunteering to participate in a new pilot?
>
> -Bill Clark
> Systems Unit
> Graduate Division
>
>
>
> -------------------------------------------------------------------------
> The following was automatically added to this message by the list server:
>
> To learn more about Micronet, including how to subscribe to or unsubscribe
> from its mailing list and how to find out about upcoming meetings, please
> visit the Micronet Web site:
>
> http://micronet.berkeley.edu
>
> Messages you send to this mailing list are public and world-viewable, and the
> list's archives can be browsed and searched on the Internet.  This means
> these messages can be viewed by (among others) your bosses, prospective
> employers, and people who have known you in the past.

 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.