IPv6 - TCP-IP

This is a discussion on IPv6 - TCP-IP ; On Aug 17, 1:20 pm, Jim Logajan wrote: > Le Chaud Lapin wrote: > > > My gut feeling is that a the the header file of a packet of Future > > Protocol need not be complex at all, ...

+ Reply to Thread
Page 2 of 2 FirstFirst 1 2
Results 21 to 31 of 31

Thread: IPv6

  1. Re: IPv6

    On Aug 17, 1:20 pm, Jim Logajan wrote:
    > Le Chaud Lapin wrote:
    >
    > > My gut feeling is that a the the header file of a packet of Future
    > > Protocol need not be complex at all, and certainly not as complex as
    > > IPv6.

    >
    > The IPv6 header per se is, IMHO, not that complicated. (Fragmentation and
    > other extension header processing on the other hand....) Which IPv6 header
    > fields do you think add needless complication that could be eliminated?


    I was mentally including the options.

    If it were me, I would strive to design the packet so that there were
    zero possibility for options. I regard options in the same way I
    regard variadic functions in C... anyone can design a function that is
    is generalized for "flexibility". This is a "design method" that is
    becoming increasingly popular in software "engineering". If you do
    not know what the form of the thing should be, write a specification
    that relieves you of having to look.

    But there is a penalty with this mode of "engineering", and it is very
    subtle. I would be wary of any specification that is underspecified,
    and that's really what options represent - an underspecification.

    -Le Chaud Lapin-





  2. Re: IPv6


    "Le Chaud Lapin" wrote in message
    news:1187414619.419459.140570@57g2000hsv.googlegro ups.com...
    > On Aug 17, 1:20 pm, Jim Logajan wrote:
    >> Le Chaud Lapin wrote:
    >>
    >> > My gut feeling is that a the the header file of a packet of Future
    >> > Protocol need not be complex at all, and certainly not as complex as
    >> > IPv6.

    >>
    >> The IPv6 header per se is, IMHO, not that complicated. (Fragmentation and
    >> other extension header processing on the other hand....) Which IPv6
    >> header
    >> fields do you think add needless complication that could be eliminated?

    >
    > I was mentally including the options.
    >
    > If it were me, I would strive to design the packet so that there were
    > zero possibility for options. I regard options in the same way I
    > regard variadic functions in C... anyone can design a function that is
    > is generalized for "flexibility". This is a "design method" that is
    > becoming increasingly popular in software "engineering". If you do
    > not know what the form of the thing should be, write a specification
    > that relieves you of having to look.
    >
    > But there is a penalty with this mode of "engineering", and it is very
    > subtle. I would be wary of any specification that is underspecified,
    > and that's really what options represent - an underspecification.


    Just like C++: underspecified, which makes it over-complex.

    John


  3. Re: IPv6

    "Le Chaud Lapin" wrote:

    > IPv6 is not clean-slate, IMO.
    >
    > If you grep for the words "backward compatible", in the IPv6 docs, you
    > will find many hits. Some authors made it at their objective to
    > change the sockets interface as little as possible, for example. I
    > have heard anyone ask the question, "What if sockets is broken too?" I
    > did read a lot of commentary that, "We should maintain backward-
    > compatibility with sockets because it is everywhere."


    Sockets is more than just IP. I'm saying that IPv6, i.e. RFC 2460, is a
    clean rewrite of IPv4, meaning RFC 791. What you are saying is that TCP
    and UDP over IPv6 are not all that different from TCP and UDP over IPv4.
    I would agree with that. However the Layer 3 stuff is very different,
    IMO.

    > IPv6 has not been accepted because failed to meet a minimum standard
    > of virtue. It is easy to blame the installed base as being
    > intransigent and not willing to deploy because it would be cost-
    > prohibited. By that is, IMO, extremely unfair to the thousands of
    > people and organizations around the world, including ISP's, who get
    > really excited at the talk of a new protocol.


    The problem is that the re-write of IP is quite total, and ISPs don't
    want to have to relearn and re-tweak everything. That seems
    understandable to me. But I wouldn't say IPv6 is finished, by any means.
    It's just that the stop-gap fixes of IPv4 have proved somewhat adequate,
    until now.

    When more devices get IP addresses and expect global connectivity, like
    any number of personal appliances, then IPv6 will likely get more
    popular very quickly.

    > People know what they like and what they do not like. They do not
    > like IPv6 as a replacement for IPv4.


    The problem I have with that comment, which I would otherwise sort of
    sympathize with, is that "people" will most likely be even more abhorent
    of a more total rewrite of Internet Protocols. Bad enough to make
    drastic layer 3 changes. If you're trying to sell drastic changes all
    the way up the stack, I wish lots of luck.

    Bert


  4. Re: IPv6

    On Aug 20, 5:25 pm, "Albert Manfredi"
    wrote:
    > "Le Chaud Lapin" wrote:
    > > People know what they like and what they do not like. They do not
    > > like IPv6 as a replacement for IPv4.

    >
    > The problem I have with that comment, which I would otherwise sort of
    > sympathize with, is that "people" will most likely be even more abhorent
    > of a more total rewrite of Internet Protocols. Bad enough to make
    > drastic layer 3 changes. If you're trying to sell drastic changes all
    > the way up the stack, I wish lots of luck.


    Let's think about that for a moment. Is that really true? Is it
    really true that people will dislike a total rewrite? It is true that
    people refuse revolutionary change simply on the basis that it is
    revolutionary? If that were true, then there would be no disruptive
    products in the history of man.

    People *will* accept the new over the old, but there is an important
    qualification that arises in each mind as the offer is considered:

    "The disparity between what I retain and what you offer must be so
    great that any pain that I suffer in switching from the old to the new
    must be worth the change."

    This qualification is true for anything.

    What happened is with IPv6 is that the...

    {increase-in-virtue}/{pain-of-adoption}

    ....ratio was too low.

    It is somewhat dishonest if we do not consider the virtue of what is
    being offered. Note by virtue, I mean the actual think, not the
    desirability of having the problems solved. We all know that we would
    like the problems solved.

    -Le Chaud Lapin-


  5. Re: IPv6


    "Le Chaud Lapin" wrote in message
    news:1188062501.895608.61880@e9g2000prf.googlegrou ps.com...
    > On Aug 20, 5:25 pm, "Albert Manfredi"
    > wrote:
    >> "Le Chaud Lapin" wrote:
    >> > People know what they like and what they do not like. They do not
    >> > like IPv6 as a replacement for IPv4.

    >>
    >> The problem I have with that comment, which I would otherwise sort of
    >> sympathize with, is that "people" will most likely be even more abhorent
    >> of a more total rewrite of Internet Protocols. Bad enough to make
    >> drastic layer 3 changes. If you're trying to sell drastic changes all
    >> the way up the stack, I wish lots of luck.

    >
    > Let's think about that for a moment. Is that really true? Is it
    > really true that people will dislike a total rewrite? It is true that
    > people refuse revolutionary change simply on the basis that it is
    > revolutionary? If that were true, then there would be no disruptive
    > products in the history of man.


    Too "deep". Have either less or more at the time you wrote that.

    > People *will* accept the new over the old, but there is an important
    > qualification that arises in each mind as the offer is considered:


    To be honest with you, I was considering contacting you for a "way out
    there" project, but now I'm not so sure. Because you seem so childlike.

    > "The disparity between what I retain and what you offer must be so
    > great that any pain that I suffer in switching from the old to the new
    > must be worth the change."
    >
    > This qualification is true for anything.


    Do yaz have to make it so damn hard to understand? Call it a compelling
    reason! Seesh.

    Jean Claude Van Damn: lighten up on the "virtue" spiel, k?

    John


  6. Re: IPv6

    On Aug 26, 12:59 am, "JohnQ"
    wrote:
    > > Let's think about that for a moment. Is that really true? Is it
    > > really true that people will dislike a total rewrite? It is true that
    > > people refuse revolutionary change simply on the basis that it is
    > > revolutionary? If that were true, then there would be no disruptive
    > > products in the history of man.

    >
    > Too "deep". Have either less or more at the time you wrote that.


    That paragraph does not seem too "deep" to me at all. Have "none" of
    what you had at the time you read it.

    > > People *will* accept the new over the old, but there is an important
    > > qualification that arises in each mind as the offer is considered:

    >
    > To be honest with you, I was considering contacting you for a "way out
    > there" project, but now I'm not so sure. Because you seem so childlike.


    Oh well. Whether I seem childlike or not, I believe what I wrote is
    true. And if you have a "way out there" project, speak up! No one is
    holding your tongue.

    > > "The disparity between what I retain and what you offer must be so
    > > great that any pain that I suffer in switching from the old to the new
    > > must be worth the change."

    >
    > > This qualification is true for anything.

    >
    > Do yaz have to make it so damn hard to understand? Call it a compelling
    > reason! Seesh.


    "Compelling reason" does not work in these kinds of arguments. The
    reason it does not work is that, with any problem, there is always a
    compelling reason. Here's an example:

    Let's say that someone took $5 billion from US government to find cure
    for pancreatic cancer, but it gave patients debilitating migraine
    headaches 22 hours / 24 hours day for rest of life, and this was
    unavoidable with the drug. The "compelling reason" would still exist -
    people do not want to die from pancreatic cancer. But the virtue of
    the solution would be in question. There might be a few people who
    would reject the solution on the "but I still won't have a life with
    massive headaches" grounds. Now take an alternate solution, one that
    actually works, but there would still be a stipulation - the patient
    would gain 10% body weight, then drop to 90%, then suffer flu-like
    symptoms 3 days straight, then break out in oozing green bed sores
    that heal only after skin therapy for a month. Many patients would
    look at this as unpleasant, but still accept the solution.

    > Jean Claude Van Damn: lighten up on the "virtue" spiel, k?


    Sorry, I will not. Virtue is what it is all about in this case.

    And "Damn" is spelled "Damme" after a city in his native Belgium.

    -Le Chaud Lapin-


  7. Re: IPv6

    On Aug 25, 1:21 pm, Le Chaud Lapin wrote:
    > On Aug 20, 5:25 pm, "Albert Manfredi"
    > wrote:
    >
    > > The problem I have with that comment, which I would otherwise sort of
    > > sympathize with, is that "people" will most likely be even more abhorent
    > > of a more total rewrite of Internet Protocols. Bad enough to make
    > > drastic layer 3 changes. If you're trying to sell drastic changes all
    > > the way up the stack, I wish lots of luck.

    >
    > Let's think about that for a moment. Is that really true?


    Yes, if you're still talking about IPv6 and not digressing into some
    other philospohical discussion that is largely irrelevant to this one.

    > Is it
    > really true that people will dislike a total rewrite? It is true that
    > people refuse revolutionary change simply on the basis that it is
    > revolutionary? If that were true, then there would be no disruptive
    > products in the history of man.


    This is what I was referring to above.

    The most compelling need for an update was, and is, the need for
    greater address space. The comprehensive rewrite of IP has instead
    introduced many other changes, like a ripple effect, that are still
    being very hotly debated. And which have yet to reach consensus.

    By the way, just parenthetically, the vast majority of what we are
    told are "disruptive products" turn out to be mostly "disruptive" in
    the imaginations of marketers and trade journal scribes, who benefit
    from such hyperbole.

    (Not to say some truly disruptive products have not been created, of
    course.)

    Bert


  8. Re: IPv6

    On Aug 26, 8:11 pm, Albert Manfredi wrote:
    > The most compelling need for an update was, and is, the need for
    > greater address space. The comprehensive rewrite of IP has instead
    > introduced many other changes, like a ripple effect, that are still
    > being very hotly debated. And which have yet to reach consensus.


    Perhaps.

    That was true in 1990. It might still be true today. But if one were
    to propose that the only change to TCP/IP be the address space, a lot
    of people, across the spectrum of users, would balk. If you're going
    to fix just the address space, why stop there? What about security?
    Mobility? Multicasting to potentially 1-billion nodes? What about
    better naming and numbering? What about an API that does not make
    coders needlessly sweat blood no matter how skilled they are? These
    are all problems that can and should be fixed.

    As far as consensus goes, I think that it is neither probable nor
    necessary. Everyone has an opinion, but it only takes a few people to
    actually find a solution.

    In the case of redoing the Internet (or any revolutionary technical
    advancement), what is required is to:

    1. Provide a fix in which the average user/programmer/ISP/venture
    capitalist/CEO/grandma/etc. can find a significant increase in virtue
    in (from their personal point of view).
    2. Introduce the fix in such away that the increase in virtue far
    exceeds the pain of acceptance.

    IPv6 failed at both these tasks taken together, and that's why it
    stands where it is today.

    -Le Chaud Lapin-


  9. Re: IPv6

    Well, almost on a lark I emailed my residential ISP asking them
    when/if IPv6 would be lit-up on residential service. My ISP responded
    by saying something that made it sound like they thought to go to IPv6
    one had to also perform a 32 to 64 bit address space migration...

    rick jones
    --
    a wide gulf separates "what if" from "if only"
    these opinions are mine, all mine; HP might not want them anyway...
    feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...

  10. Re: IPv6

    On Aug 27, 1:32 pm, Rick Jones wrote:
    > Well, almost on a lark I emailed my residential ISP asking them
    > when/if IPv6 would be lit-up on residential service. My ISP responded
    > by saying something that made it sound like they thought to go to IPv6
    > one had to also perform a 32 to 64 bit address space migration...


    LOL. Did he mean the bit-width of CPU registers/address bus?

    -Le Chaud Lapin-


  11. Re: IPv6

    Le Chaud Lapin wrote:
    > On Aug 27, 1:32 pm, Rick Jones wrote:
    > > Well, almost on a lark I emailed my residential ISP asking them
    > > when/if IPv6 would be lit-up on residential service. My ISP responded
    > > by saying something that made it sound like they thought to go to IPv6
    > > one had to also perform a 32 to 64 bit address space migration...


    > LOL. Did he mean the bit-width of CPU registers/address bus?


    Well, I was reading between the lines s/he was given to repeat, but I
    think that there was indeed some confustion about 32-bit IPv4 vs
    128-bit IPv6 and 32-bit versus 64-bit pointers/whatnot.

    I replied pointing-out how application changes to enable IPv6 didn't
    mean migrating from an ILP32 model and how it wasn't all that
    difficult - the support person was repeating that IPv6 migration would
    be like rewriting applications from scratch - and am anxious to see
    what response if any ensues.

    Just another intesting tidbit on the topic of IPv6 deployment...

    rick jones
    --
    Wisdom Teeth are impacted, people are affected by the effects of events.
    these opinions are mine, all mine; HP might not want them anyway...
    feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...

+ Reply to Thread
Page 2 of 2 FirstFirst 1 2