Exceed onDemand versus VNC - Xwindows

This is a discussion on Exceed onDemand versus VNC - Xwindows ; On Thu, 30 Sep 2004 01:48:12 +0000, sinister wrote: > "Richard Santink" wrote in message > news:415B4A0C.3000005@nortel.ca... [snippage] > Both X-Win32 (which worked fine from a PC to the Unix host over the > local ethernet at work) and Exceed ...

+ Reply to Thread
Page 2 of 2 FirstFirst 1 2
Results 21 to 27 of 27

Thread: Exceed onDemand versus VNC

  1. Re: Exceed onDemand versus VNC

    On Thu, 30 Sep 2004 01:48:12 +0000, sinister wrote:
    > "Richard Santink" wrote in message
    > news:415B4A0C.3000005@nortel.ca...

    [snippage]
    > Both X-Win32 (which worked fine from a PC to the Unix host over the
    > local ethernet at work) and Exceed were so slow as to be useless. I did
    > a little research, and it seems that bandwidth isn't the only issue, but
    > also "network latency." X does some kind of synchronization or
    > something, whereby packets are sent back and forth. It might be OK over
    > an ethernet, where I suppose the circuit time is on the order of 1 ms
    > or less, but over something like DSL (with the circuit time being on the
    > order of tens of ms), it'll just kill you.


    ISTR a buddy of mine used X over serial dialup years ago. It is very slow.
    There was a way to improve it by some kind of compression (?) on the
    serial transmission, or was that some "middleware" that did something
    special with polling and/or synchronization (as you put it?). It never was
    great, but it became usable. I'm not any kind of expert in this area. If
    interested (and if no one else here recalls the method) I can try to get
    that info (but it may take time) from my buddy of years ago.

    >> As for their use of the terms 'server' and 'client'...

    >
    > ...VNC...Exceed... has the same nomenclature (i.e., the one
    > opposite of the usual X nomenclature).


    I think the confusion arises because most of the punters think in terms of
    boxes, and not services. The X server provides a display service,
    therefore it obviously runs on the computer/device with the display
    attached. Services can be provided all over the place (including from
    "workstations"), so it is meaningless to talk of "THE server" (box).
    Is this all legacy from PCs attached to a (Corvus?) file/print server?

    --
    Juhan Leemet
    Logicognosis, Inc.


  2. Re: Exceed onDemand versus VNC

    Kyler Laird writes:

    > Yup, bandwidth/latency problems are why LBX was developed.
    > http://studenthelp.itee.uq.edu.au/remote/lbx.html


    LBX isn't that useful according to Keith Packard:

    http://keithp.com/~keithp/talks/lbxpost/paper.html

    Mr. Packard did some analysis of X network performance and generally
    concluded that using SSH compression is as good as, or better than,
    using LBX:

    http://keithp.com/~keithp/talks/usenix2003/

    The main issue with SSH is it can increase latency on high bandwidth
    links. From the conclusions:

    Performance using LBX over the highest latency links is no better
    than SSH. Performance over broadband network of the protocol
    without compression is only 25 percent worse than LBX. LBX does
    not solve either the authentication and security problems that
    SSH solves. We saw little evidence of LBX ever helping. At least
    as implemented, LBX looks to have been a bad idea

    --
    David Magda , http://www.magda.ca/
    Because the innovator has for enemies all those who have done well under
    the old conditions, and lukewarm defenders in those who may do well
    under the new. -- Niccolo Machiavelli, _The Prince_, Chapter VI

  3. Re: Exceed onDemand versus VNC

    After takin a swig o' Arrakan spice grog, David Magda belched out:
    > We saw little evidence of LBX ever helping. At least as
    > implemented, LBX looks to have been a bad idea


    That strikes me as overstating things in a pretty mean fashion.

    LBX may have turned out to be unsuccessful, but you often have to go
    down some blind paths that _weren't_ necessarily "bad ideas" in order
    to find the right path.

    We solved a performance problem yesterday (almost a month after the
    initial report) that was very much like that; there was a very
    particular problem involving a runaway monitoring process.

    We couldn't _find_ it until we went down several "blind paths" to rule
    out things as NOT being broken.

    I see no reason to have expected LBX to be a "bad idea," a priori. It
    turns out that it wasn't as good as SSH tunnelling, but we couldn't
    really know that before both were tried.
    --
    wm(X,Y):-write(X),write('@'),write(Y). wm('cbbrowne','acm.org').
    http://www3.sympatico.ca/cbbrowne/postgresql.html
    "Be humble. A lot happened before you were born." - Life's Little
    Instruction Book

  4. Re: Exceed onDemand versus VNC


    "Kyler Laird" wrote in message
    news:7tft22-067.ln1@lairds.us...
    > "sinister" writes:
    >
    >>The particular Exceed product I'm looking at is Exceed onDemand.

    >
    > It looks like they're providing some sort of low-bandwidth proxying.
    > In this case, we're no longer talking about X. It's behaving *just*
    > like VNC.


    OK. Though in its favor, I'm pretty sure it includes encryption, which VNC
    doesn't. Of course, you'd think the guy would just set up ssh, but no...

    >
    > Heck, if you want to go with a proprietary solution and get suspend/
    > resume functionality, you might as well use something like XVision
    > Eclipse.
    > http://www.tarantella.com/products/v...ds.html#resume


    One thing that's not clear to me (I'm not referring to your posts, but to
    the terms as generally used): Does suspend/resume allow the apps you're
    using to continue working? E.g. I often set up software sessions that
    involve lots of number crunching and want to let them run on the UNIX host
    at work *and* turn my home PC off. I can easily do it with VNC; just
    disconnect the client on my home PC. But I'm not sure about Exceed.

    >
    >>Both X-Win32 (which worked fine from a PC to the Unix host over the local
    >>ethernet at work) and Exceed were so slow as to be useless. I did a
    >>little
    >>research, and it seems that bandwidth isn't the only issue, but also
    >>"network latency." X does some kind of synchronization or something,
    >>whereby packets are sent back and forth. It might be OK over an ethernet,
    >>where I suppose the circuit time is on the order of 1 ms or less, but
    >>over
    >>something like DSL (with the circuit time being on the order of tens of
    >>ms),
    >>it'll just kill you.

    >
    > Yup, bandwidth/latency problems are why LBX was developed.
    > http://studenthelp.itee.uq.edu.au/remote/lbx.html


    OK.

    >
    >>> As for their use of the terms 'server' and 'client' I'm still a bit
    >>> foggy..

    >
    >>In this case it's probably because their "server" is serving up a special
    >>protocol that enables communication over a slower network (e.g. DSL) to be
    >>do-able.

    >
    > Agreed.


    OK.

    >
    >>VNC (which I actually think is fine, but my sysadmin is pushing
    >>Exceed...) has the same nomenclature (i.e., the one opposite of the usual
    >>X
    >>nomenclature).

    >
    > So he wants a less-functional (costly) proprietary solution to X proxying
    > instead of an Open solution which does everything desired and more? I'd
    > expect that from management, not a sysadmin.


    Well, he has a propensity for liking proprietary solutions.

    The stupidest thing is when I got to the job a few months ago, everyone who
    had to run stuff on the Sun Solaris server had a SunRay or SunBlade. *And*
    they had their own PC (which they'll want in any case). I thought, "why not
    just use VNC from the PC?"

    It's a real pain the a** his way, because you have to constantly switch
    mouse and keyboard. (And Sun and PC keyboards are slightly differently laid
    out.) And as you point out his solution is more expensive.

    IIRC he was pretty uptight about people connected with VNC with no
    encryption. I had thought that the solution to that would be set up ssh,
    but he was too lazy to do that.

    Past tense above, because he was drummed out of the office Friday for some
    kind of no-no. There had been no sudden failure (like lost data), just the
    same continuing level of (in)competence. Some kind of non-technical ethics
    thing got his butt canned.

    Now we have to look for a new one...

    Thanks for your replies,

    S

    >
    > --kyler




  5. Re: Exceed onDemand versus VNC


    "David Magda" wrote in message
    news:86zn358ajl.fsf@number6.magda.ca...
    > Kyler Laird writes:
    >
    >> Yup, bandwidth/latency problems are why LBX was developed.
    >> http://studenthelp.itee.uq.edu.au/remote/lbx.html

    >
    > LBX isn't that useful according to Keith Packard:
    >
    > http://keithp.com/~keithp/talks/lbxpost/paper.html
    >
    > Mr. Packard did some analysis of X network performance and generally
    > concluded that using SSH compression is as good as, or better than,
    > using LBX:
    >
    > http://keithp.com/~keithp/talks/usenix2003/
    >
    > The main issue with SSH is it can increase latency on high bandwidth
    > links. From the conclusions:


    On *high* bandwidth links...

    How much effect does SSH have on latency over something like DSL or cable?

    >
    > Performance using LBX over the highest latency links is no better
    > than SSH. Performance over broadband network of the protocol
    > without compression is only 25 percent worse than LBX. LBX does
    > not solve either the authentication and security problems that
    > SSH solves. We saw little evidence of LBX ever helping. At least
    > as implemented, LBX looks to have been a bad idea
    >
    > --
    > David Magda , http://www.magda.ca/
    > Because the innovator has for enemies all those who have done well under
    > the old conditions, and lukewarm defenders in those who may do well
    > under the new. -- Niccolo Machiavelli, _The Prince_, Chapter VI




  6. Re: Exceed onDemand versus VNC


    "Juhan Leemet" wrote in message
    newsan.2004.09.30.16.07.42.292260@logicognosis.com...
    > On Thu, 30 Sep 2004 01:48:12 +0000, sinister wrote:
    >> "Richard Santink" wrote in message
    >> news:415B4A0C.3000005@nortel.ca...

    > [snippage]
    >> Both X-Win32 (which worked fine from a PC to the Unix host over the
    >> local ethernet at work) and Exceed were so slow as to be useless. I did
    >> a little research, and it seems that bandwidth isn't the only issue, but
    >> also "network latency." X does some kind of synchronization or
    >> something, whereby packets are sent back and forth. It might be OK over
    >> an ethernet, where I suppose the circuit time is on the order of 1 ms
    >> or less, but over something like DSL (with the circuit time being on the
    >> order of tens of ms), it'll just kill you.

    >
    > ISTR a buddy of mine used X over serial dialup years ago. It is very slow.
    > There was a way to improve it by some kind of compression (?) on the
    > serial transmission, or was that some "middleware" that did something
    > special with polling and/or synchronization (as you put it?). It never was
    > great, but it became usable. I'm not any kind of expert in this area. If
    > interested (and if no one else here recalls the method) I can try to get
    > that info (but it may take time) from my buddy of years ago.


    VNC is very good for this. I started the whole thread because my sysadmin,
    who likes to push proprietary services, wanted us to use Exceed onDemand
    rather than VNC.

    >
    >>> As for their use of the terms 'server' and 'client'...

    >>
    >> ...VNC...Exceed... has the same nomenclature (i.e., the one
    >> opposite of the usual X nomenclature).

    >
    > I think the confusion arises because most of the punters think in terms of
    > boxes, and not services. The X server provides a display service,
    > therefore it obviously runs on the computer/device with the display
    > attached. Services can be provided all over the place (including from
    > "workstations"), so it is meaningless to talk of "THE server" (box).


    Right.

    > Is this all legacy from PCs attached to a (Corvus?) file/print server?


    As the OP, it concerns connecting from PCs (both over high-bandwidth etheret
    and over DSL/cable) to Solaris.

    > --
    > Juhan Leemet
    > Logicognosis, Inc.
    >




  7. Re: Exceed onDemand versus VNC


    "Kyler Laird" wrote in message
    news:knet22-067.ln1@lairds.us...
    > "sinister" writes:
    >
    >>> The X terminology is confusing. I'd write it like this.
    >>> VNC: 3 pieces of software (VNC server, VNC client, applications)
    >>> Exceed plus xmove: 3 pieces of software (xmove, Exceed server,
    >>> applications)

    >
    >>You forgot the Exceed client.

    >
    > I used the term "applications" instead of "X clients" in order to
    > be more parallel with the wording in the VNC line.
    >
    > I understand that saying "Exceed server" instead of "X server" might
    > be contrary to their terminology. I probably should have written
    > "Exceed client."


    OK. Sorry for being testy.

    >
    > --kyler




+ Reply to Thread
Page 2 of 2 FirstFirst 1 2