471,109 Members | 1,491 Online
Bytes | Software Development & Data Engineering Community
Post +

Home Posts Topics Members FAQ

Join Bytes to post your question to a community of 471,109 software developers and data experts.

Rewrite TCP/IP From Scratch

Before I begin, I must emphasize the word "Scratch".

Normally when I make this proposal, I have to go through 7 or 8
iterations until my victim finally realizes that when I say "scratch",
I do mean "scratch." So perhaps the title of this post should be
"Design A New Networking Protocol From Scratch With Near-Zero Regard
For TCP/IP." No attempt shall be made to synthesize the engine via
incremental modification of existing protocols (as was done with
IPv6). We will necessarily make a clean break from all things TCP,
UDP, IP, DCE, CORBA, DCOM, .NET, RPC, SSL, sockets, ARP, RARP, SNMP,
CIDR, NAT, EGP, RIP, DNS, etc. If any legacy elements are recognized
in the new protocol, it shall be those that are fundamentally
persistent, like the end-to-end model and retransmission as a
mechanism to overcome errors.

I would like to propose that a group of us band to together to find a
virtuous framework to the problem that can be broadly defined as
"Distributed Communication". This proposal is relevant to C++ because
its reference implementation will be in C++.

Specifically, I propose that we create a new full-featured networking
engine that facilitates distributed communication between machines on
a global network. This engine will provide mechanisms for naming,
numbering, addressing, security, mobility, multicasting, and highly
elegant interprocess communication. A library shall be created to
allow third-party developers to write applications against the new
DCE.

After this new engine has been created, we will choose, from among
several "killer" applications, one that demonstrates both the
potential and necessity of the new distributed communications engine,
and implement a prototype of that application against the new engine.

All software shall be written in C++ and made strictly portable
(greater than 95% inter-OS compilability with no allowance for
#ifdef's).

If I did not feel that I had the technical insight to guide such a
project to success, I would not have made the proposal.

This has been a dream of mine for quite a while. If I lost both my
hands, I would continue to type with my teeth to see that it was done.
I am firmly comitted.

I hope to appeal to the insight, creativity, ambition, perceptivity,
and intelligence in all of you, as it would be much more enjoyable to
not do it alone.

I know some of the best C++ programmers in the world lurk in this
group. It is to you I make this appeal. If you have an opinion one
way or another, I would like to hear it, here, or via email.

Best Regards,

-J. C. Jones-
Research Scientist
Jul 22 '05 #1
8 5442
On 13 Jul 2004 15:15:57 -0700, un*****************@yahoo.com (Le Chaud
Lapin) wrote in comp.lang.c++:
Before I begin, I must emphasize the word "Scratch".

Normally when I make this proposal, I have to go through 7 or 8
iterations until my victim finally realizes that when I say "scratch",
I do mean "scratch." So perhaps the title of this post should be
"Design A New Networking Protocol From Scratch With Near-Zero Regard
For TCP/IP."


[snip]

What does any of this have to do with the C++ language? What
personality flaw caused you to post this here, where it is 100%
off-topic?

--
Jack Klein
Home: http://JK-Technology.Com
FAQs for
comp.lang.c http://www.eskimo.com/~scs/C-faq/top.html
comp.lang.c++ http://www.parashift.com/c++-faq-lite/
alt.comp.lang.learn.c-c++
http://www.contrib.andrew.cmu.edu/~a...FAQ-acllc.html
Jul 22 '05 #2
un*****************@yahoo.com (Le Chaud Lapin) wrote in message news:<fc**************************@posting.google. com>...

[snip - no TCP]
Specifically, I propose that we create a new full-featured networking
engine that facilitates distributed communication between machines on
a global network. This engine will provide mechanisms for naming,
numbering, addressing, security, mobility, multicasting, and highly
elegant interprocess communication. A library shall be created to
allow third-party developers to write applications against the new
DCE.


Fortunately for us, this task is already accomplished. It's called 9p2000.

/david
Jul 22 '05 #3
David Rubin wrote:
Le Chaud Lapin wrote:

[snip - no TCP]
Specifically, I propose that we create a new full-featured networking
engine that facilitates distributed communication between machines on
a global network. This engine will provide mechanisms for naming,
numbering, addressing, security, mobility, multicasting, and highly
elegant interprocess communication. A library shall be created to
allow third-party developers to write applications against the new
DCE.


Fortunately for us, this task is already accomplished. It's called 9p2000.


Please tell us it's Spam- and virus-proof.

--
Phlip
http://industrialxp.org/community/bi...UserInterfaces
Jul 22 '05 #4
Jack Klein wrote:
Jack Klein
Home: http://JK-Technology.Com


Nice site, however it would be better if you fixed these two to be ISO C++:
http://jk-technology.com/c/code/comments.html

http://jk-technology.com/c/code/cpptohtml.html


Regards,

Ioannis Vranos

http://www23.brinkster.com/noicys
Jul 22 '05 #5
"Phlip" <ph*******@yahoo.com> wrote in message
news:vA*************@newssvr19.news.prodigy.com...
David Rubin wrote:
Le Chaud Lapin wrote:

[snip - no TCP]
Specifically, I propose that we create a new full-featured networking
engine that facilitates distributed communication between machines on
a global network. This engine will provide mechanisms for naming,
numbering, addressing, security, mobility, multicasting, and highly
elegant interprocess communication. A library shall be created to
allow third-party developers to write applications against the new
DCE.
Fortunately for us, this task is already accomplished. It's called

9p2000.
Please tell us it's Spam- and virus-proof.


Just like in the long run, nothing is proof to cockroaches, nothing is proof
to spam and viruses. To make a system that would be would require the
elimination of some of the features most necessary in an internet protocol.
For instance, the independent creation of accounts for e-mail or news would
have to die to eliminate spam, because the only way to hold people
accountable for what they write is to centralize account naming and cause it
to require verifiable data about who a potential account holder is. To
eliminate viruses, one would have to eliminate a person's ability to send
any kind of binary. Actually, the only way to truly and permanently protect
against viruses is to remove all input devices from your computer.
Jul 22 '05 #6
Jack Klein <ja*******@spamcop.net> wrote in message news:<ad********************************@4ax.com>. ..
On 13 Jul 2004 15:15:57 -0700, un*****************@yahoo.com (Le Chaud
What does any of this have to do with the C++ language? What
personality flaw caused you to post this here, where it is 100%
off-topic?


Gee Jack, I do not know. Maybe you could fly your personal
psychiatrist over to France and let me borrow him/her long enough so
that I might provide a suitable answer to your (second) question.

In the meantime, I think I will post at will, as I feel that this has
some interest to C++ programmers, given that most of are beginning to
get into distributed development if we have not already.

-Chaud Lapin-
Jul 22 '05 #7
Le Chaud Lapin wrote:
Gee Jack, I do not know. Maybe you could fly your personal
psychiatrist over to France and let me borrow him/her long enough so
that I might provide a suitable answer to your (second) question.

In the meantime, I think I will post at will, as I feel that this has
some interest to C++ programmers, given that most of are beginning to
get into distributed development if we have not already.
Folks interested in network protocols can easily find their newsgroups.
Newsgroups work best when people qualified to answer on a given group don't
find too many irrelevant questions to waste their time.

However, I won't read your posts here or there. >Plonk<
-Chaud Lapin-


Cool handle, though!

--
Phlip
http://industrialxp.org/community/bi...UserInterfaces
Jul 22 '05 #8
"Phlip" <ph*******@yahoo.com> wrote in message news:<Fi****************@newssvr33.news.prodigy.co m>...
Le Chaud Lapin wrote:
Gee Jack, I do not know. Maybe you could fly your personal
psychiatrist over to France and let me borrow him/her long enough so
that I might provide a suitable answer to your (second) question.

In the meantime, I think I will post at will, as I feel that this has
some interest to C++ programmers, given that most of are beginning to
get into distributed development if we have not already.
Folks interested in network protocols can easily find their newsgroups.
Newsgroups work best when people qualified to answer on a given group don't
find too many irrelevant questions to waste their time.

However, I won't read your posts here or there. >Plonk<


Looking at some of the posts you have made, I think I will reciprocate.
whatever word u find appropriate<


-Chaud Lapin-
Jul 22 '05 #9

This discussion thread is closed

Replies have been disabled for this discussion.

Similar topics

reply views Thread by Link | last post: by
1 post views Thread by darrel | last post: by
1 post views Thread by baroque Chou | last post: by
14 posts views Thread by Stan Canepa | last post: by
8 posts views Thread by white lightning | last post: by
10 posts views Thread by not_a_commie | last post: by

By using Bytes.com and it's services, you agree to our Privacy Policy and Terms of Use.

To disable or enable advertisements and analytics tracking please visit the manage ads & tracking page.