[THIN] Re: Pano VDI Solution
- From: "Steve Greenberg" <steveg@xxxxxxxxxxxxxx>
- To: <thin@xxxxxxxxxxxxx>
- Date: Sun, 8 Jun 2008 13:07:11 -0700
Common reply to Rick and Mike-
These are both great comments. When I take this feedback and put it in
context of 15 years of doing this remote computing stuff, I am seeing a
covergange up head. i.e. A total device that is at once the Net2Display
"decoder" and a thin client and a rich PC. That is, a solid state device
that has the intelligence of low level network optimization, remote display
client and local decoders/media clients as well as a Calista like
client/graphics server.
With practical solid state drives just around the corner, perhaps the new
thing will really be all the old things wrapped up into one smart device. It
is all solid state and seamlessly mixes local execution with remote
displays-- when you walk up, login or put in your tiny 2TB solid state key,
it becomes everything you need on the fly. Let's say you have LAN
connectivity, then everything is remoted. Let's say you have mediocre WAN,
it caches and downloads those elements that can't run well interactively.
Let's say that you are on a high bandwidth but high latency connection like
satellite or across the planet on a WAN, then it would actually stream a
virtual machine down to the local device to run and the files you need...
To guide this powerful solid state device we would need an open format to
communicate the conditions and requirements of each element of the session,
something like XML for remoting, i.e. XDL- Extensible Display Language, a
way of telling the device you are sitting exactly how to handle each
application, rich media stream, voice conversation, remote machine, virtual
machines, etc..
Just dreaming out loud, but it makes sense to me!!
Steve Greenberg
Thin Client Computing
34522 N. Scottsdale Rd D8453
Scottsdale, AZ 85266
(602) 432-8649
www.thinclient.net
steveg@xxxxxxxxxxxxxx
_____
From: thin-bounce@xxxxxxxxxxxxx [mailto:thin-bounce@xxxxxxxxxxxxx] On Behalf
Of Mike Hancock
Sent: Sunday, June 08, 2008 9:56 AM
To: thin@xxxxxxxxxxxxx
Subject: [THIN] Re: Pano VDI Solution
Couldn't companies like Citrix move more of the intelligence down the stack
and into the networking layers by leveraging existing products like
WANscaler instead of depending on the client device for the buffering? In
theory, a PC, thin client or whatever, and any software on those devices
should never have to know if it is getting data from a LAN or a WAN.
I know Cisco is moving toward taking the WAN and emulating a LAN with more
intelligent networking, prioritization, security and caching and pitting it
onto their routers and switches.
Just a thought
Mike
From: thin-bounce@xxxxxxxxxxxxx [mailto:thin-bounce@xxxxxxxxxxxxx] On Behalf
Of Rick Mack
Sent: Sunday, June 08, 2008 6:25 AM
To: thin@xxxxxxxxxxxxx
Subject: [THIN] Re: Pano VDI Solution
Hi Steve,
The Net2Display standard is real and is set to be ratified later this year.
Considering IBM's contribution to Net2Display I don't think I'd lose any
money on a bet that they'll be one of the first implementers, but initially
for blade PCs only. On the thin client front, Devon IT might very well be
first out of the blocks and considering the relations between VMware and
Microsoft, it wouldn't be all that surprising to see Net2Display support in
VDM3.
I guess I've got to qualify my comments with regards to Net2Display and
RDP/ICA because life is never totally simple.
If we think about what Panologic and Teradic are doing at the moment then
it's likely that we will have no operating system as such on the Net2Display
thin client devices, at least for LAN-connected devices (see comments about
WAN below). That's pretty profound and will mean that we will finally have
thin client devices are much less expensive than a PC with no compromises in
the area of graphics, USB support and multimedia. Imagine what that will do
for thin client-based technologies in education.
I think that Net2Display will be the killer protocol for LAN conencted thin
clients and will very likely displace RDP and ICA. But multimedia support
and WAN connectivity unfortunately raises some special challenges that will
redefine Net2Display capabilities.
If we have a look at ICA as an example, you have separate audio and video
channels and run them at different priorities. Throw in different
compression technologies for audio and video as well then synchronisation
of audio and video becomes hit or miss, mostly the latter.
The only way to handle multimedia redirection effectively is by redirecting
the multimedia content to the appropriate player on a thin client device
with enough buffereing to deal with bandwidth variability. That
unfortunately results in increasing the complexity of the operating system
at the client end with a multitude of codecs, players and IE plugins. If
you've got embedded XP or XP/Vista running on all your thin client machines
just to support multimedia, at least some of the promised TCO savings
compared to managing traditional fat clients get swallowed up managing the
thin client device operating systems.
The challenge for the Net2Display protocol is that playing multimedia
content on the thin client device is still the most efficient way to do it
in terms of bandwidth utilization even with stuff like lossy compression.
That doesn't matter on a LAN or where you've got lots of bandwidth, but it
does matter on a high-latency and slow or variable bandwidth WAN
connections.
We also haven't addressed latency effects that need stuff like a local text
echo, local cursor, additional TCP/IP optimization and so on to make an
acceptable user experience. That's going to require somewhat more
intelligence at the thin client end and means that at the least we will need
a minimal latency-aware operating system that supports multimedia players.
So for WAN connected thin clients, Net2Display probably isn't going to
provide the functionality needed unless it's a piece of client software
running on an operating system that supports multimedia redirection and all
the other latency reduction stuff. That sounds a lot like what we've got
already with ICA and RDP.
I'd better finish rambling now with a bit of a summary.
Net2Display has a very good chance of dominating the LAN-connected thin
client market. While I'm not discounting the effects of the Callista
technology on RDP, Microsoft is going to have to do something really
interesting on the thin client o.s. licensing side to make RDP competitive
in that area.
The outcome in the WAN connectivity area is still wide open. Provision
Network's enhanced RDP is catching up with ICA and unless Citrix do
something really smart like WANScaler based ICA bitmap caching to give ICA a
profound advantage, the ICA/RDP comparison is becoming irrelevant. Combine
Net2Display with latency reduction and multimedia redirection and you've got
a LAN and WAN solution that actually has a good chance to replace RDP and
ICA.
regards,
Rick
--
Ulrich Mack
Quest Software
Provision Networks Division
Other related posts:
