Laconsults | 1 Oct 2005 21:43
Picon
Favicon

Re: FREE LICENSE BUNDLING Historic Significant Next Generation TCP onto IwIP ...

In a message dated 01/10/2005 20:39:39 GMT Daylight Time, Laconsults writes:



RE : PATENT LICENSE BUNDLING Historic Significant Next Generation TCP patent technology breakthroughs onto IwIP TCP  : not a single packet/s transmitted ever gets dropped &all packets each arrive with PSTN transmissions quality worldwide



TCP researchers Interested in collaborating to free license bundle Historic Significant Next Generation TCP onto  IwIP TCP , please email patent <at> iwxchange.com implementation details &guide ( just simple few lines of modifications required on existing WATTCP source codes , immeditely enabling not a single packet/s transmitted to be ever dropped &all packets each arrive at destination with PSTN transmissions quality worldwide )

first visit http://iwxchange.com for some earlier developed technologies details, or Google search ' IWXChange '

Look forward to hear from you to begin collaborations.

Thank you,
EMail : laconsults <at> aol.com



MOST RECENT BACKGROUND ARTICLES
================================


Sawtoth vs Pause : TCP Accelerator SAN Storage Area Network New Patent Groundbreaking Technology

08-20-05 10:50 PM



Please Note : this next pending NextGenTCP release is now increment
deployable &TCP friendly


the next release of one-click self installing NextGen TCP upgrade
software is increment deployable on any Linux/ Window PC , with
immediate immense benefits even if yours is the only PC worldwide using
NextGen TCP : moreover where subsequently  there exists a majority of
PCS within any geographical Internet subset/s using NextGenTCP, the
data transmissions within the subset/s could become same as PSTN
transmissions quality even for other non-adopters !


appended here some backgrounds on our increment deployable TCP friendly
NextGen TCP patent technologies, this could generates tens of billions
revenues together with corporations with pre-eminent innovative
marketplace reputations, like hand-in-glove..

see also , for the published PCT material related to the technology (
Note : there are further as yet unpublished patent filings on 100% SAN
link utiklisation &other improvements :

1 IMMEDIATE READY IMPLEMENTATION OF VIRTUALLY CONGESTION FREE
GUARANTEED SERVICE CAPABLE NETWORK

Publication info: WO2005053265 - 2005-06-09

if some tech corporations were asked a year ago to write a Million
dollar research cheque finding a solution to the TCP Accelerator
issues, they will be mightiliy please if guaranteed any sort of results
even if not know what the solution might even be.


The Swedish http://blue2net.com  corporation's CEO Founder recently
paid one of our Directors to Stockholm, &our NextGenTCP Video was
successfully reproduced first time attempted on their own provided
LapTops

he was there in person to see the first time successful presentations
on their own provided laptops.

he thought we could be Moses &push us for same NextGenTCP that works
immediately over public external Internet so he can use PSTN quality
VoIP over external Internet !  Ha Ha !

strange, didnt think so at the time, but we are now getting there
....starting with this most recent  constant 100% TCP link utilisation
throughput  over external Internet patent... could  further be adapted
into PSTN quality real time over external Internet in this next pending
software release...



Regards,
IWXChange
http://iwxchange.com
EMail : patent <at> iwxchange.com




laconsults <at> aol.com wrote:
>TCP Accelerator SAN Storage Area Network New
>Patent Groundbreaking Technology company : making existing persistent
>vexing throughputs problems on LFN long distance high latencies /
>Inter-Continental Public Internet SAN Storage Data Transfers now ' a
>complete past history '
>
>
>the graph here looks ' very beautiful ' .... like money written all
>over it , when It was first successfully produced in tests (
>http://members.aol.com/laconsults/SquareWave  )
>
>
>Regards,
>IWXChange
>Swiss Cottage, UK
>http://IWXChange.com
>EMail : IWXCha... <at> aol.com
>
>
>a new breakthrough is achieved with new TCP ' SQUARE-WAVE ' constant
>100% link's bandwidth utilisation throughout( see Ethereal IO-Graph
>http://members.aol.com/laconsults/SquareWave  )
>
>
>we are now focused with our developers to adapt our technology for some
>high profile organisations here in UK.
>
>
>at present costs big companies US$50K -100K deploying latest state of
>art TCP accelerator software ' which werent even all that good in
>physical link's bandwidth utilisations improvements ( awfull ! but even
>worse without ) . This market is worth tens of billions annually.
>
>
>test results on the adapted software's ( ie adapting the now well
>established NextGenTCP's ' Pause ' new breakthrough technology ) : very
>effective ' constant near 100% throughput ' as is seen in the ethereal
>IO-Graph's ' Square Wave Signal Form compared to usual ' zagged see-saw
>' forms constantly fluctuating between 0% &100% seriously wasting
>bandwidths.
>
>
>This Patent now totally removes all ' public Internet congestions drops
>' and BER ( physical transmissions bit error rates , especially high in
>Satellites/ Mobile 3G ) and all RTT latencies effects . Typical Asia
>Internet Drop rates is 40% ! North America 5% , even 1% public Internet
>drop rates equates to 50% bandwidth utilisation / throughputs
>reductions. See http://internettrafficreport.com  ( UK to Singapore
>typical droprates is 5% hence not even meaningful trying to do branch
>data back-up to Singapore at all , UK to Japan typically 10% drop rates
>) . The existing well documented persistent vexing problems associated
>with throughputs problems on all long distance high latencies Long Fat
>Network ( LFN ) , has now been made ' history ' with this
>groundbreaking new patent technology.
>
>SAN providers/ innovators interested in collaboration developments
>please email us for complete technology disclosures.
>
>The industrial strength commercial software should be ready soon , no
>other TCP accelerators can even come close ( &recently there were
>several TCP accelerator start-ups acquired by big players for billions
>dollars each , &they weren't even in our league )
>










_______________________________________________
lwip-devel mailing list
lwip-devel <at> nongnu.org
http://lists.nongnu.org/mailman/listinfo/lwip-devel
Sathya Thammanur | 4 Oct 2005 16:52
Picon

Re: lwip memory on a xilinx ml403 dev board

Hi Kyle,
What about the stack and heap size? Are you quoting the code foot print here or is it the complete program size? Are you compiling your program with the right optmization level ? Do give in more information and I can see what is going on.

Sathya


On 9/30/05, kyle treleaven <nightfender-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
Hi all,

I'm trying to use the Xilinx-provided port of lwIP on a PPC405
(embedded) running the "standalone" OS-- ie. no OS.  I'm using the raw
API, and I've simply compiled a bit of tutorial code (a small echo
server) provided in an appnote by Xilinx-- I'm sure a lot of people
have run the same drill.  But for some reason my compiles keep running
up in the 800K range, which is way way way too big for my poor little
fpga.  Plus it seems ridiculous when compared to the "tens of
kilobytes" that everybody else claims on other platforms.  My mem/memp
settings are (the defaults):

MEM_SIZE: 8192
MEMP_NUM_PBUF: 16
MEMP_NUM_UDP_PCB: 5
MEMP_NUM_TCP_PCB: 5
MEMP_NUM_TCP_PCB_LISTEN: 5
MEMP_NUM_TCP_SEG: 255

I'm under the impression that these are the only settings that matter
for raw api use.

Has anybody else had this problem? Anybody know what's going on?

Kyle


_______________________________________________
lwip-users mailing list
lwip-users-qX2TKyscuCcdnm+yROfE0A@public.gmane.org
http://lists.nongnu.org/mailman/listinfo/lwip-users

_______________________________________________
lwip-users mailing list
lwip-users@...
http://lists.nongnu.org/mailman/listinfo/lwip-users
Vincent Bruyere | 4 Oct 2005 18:39
Picon
Favicon

fd_sets building

Good evening,

I am working on an embedded MMS project with modem and
PPP support.
I am using lwip1.1.0 stack.
I have several problems when receiving files bigger
than 3 KBytes.
The socket is often blocking to get data so i wonder
if the socket select management is right.

Does anybody know how to manage fd_sets building
before running select, for a non UNIX OS ?
ie is it necessary to use FD_ZERO, FD_SET, FD_ISSET
like commands on real-time OS ?

Thanks in advance and regards,

Vincent BRUYERE

	

	
		
___________________________________________________________________________ 
Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger 
Téléchargez cette version sur http://fr.messenger.yahoo.com
Bertrik Sikken | 4 Oct 2005 20:54
Picon

Re: lwip memory on a xilinx ml403 dev board


kyle treleaven wrote:
> Hi all,
> 
> I'm trying to use the Xilinx-provided port of lwIP on a PPC405
> (embedded) running the "standalone" OS-- ie. no OS.  I'm using the raw
> API, and I've simply compiled a bit of tutorial code (a small echo
> server) provided in an appnote by Xilinx-- I'm sure a lot of people
> have run the same drill.  But for some reason my compiles keep running
> up in the 800K range, which is way way way too big for my poor little
> fpga.  Plus it seems ridiculous when compared to the "tens of
> kilobytes" that everybody else claims on other platforms.  My mem/memp
> settings are (the defaults):
> 
> MEM_SIZE: 8192
> MEMP_NUM_PBUF: 16
> MEMP_NUM_UDP_PCB: 5
> MEMP_NUM_TCP_PCB: 5
> MEMP_NUM_TCP_PCB_LISTEN: 5
> MEMP_NUM_TCP_SEG: 255
> 
> I'm under the impression that these are the only settings that matter
> for raw api use.
> 
> Has anybody else had this problem? Anybody know what's going on?

What kind of executable are you compiling?
(is it an executable? or is it some kind of ROM image, or FPGA image?)

If possible, try to convince the linker to create a map file for you.
The map file should tell you exactly where the memory is going.

All the best,
Bertrik
Atte Kojo | 5 Oct 2005 08:01
Picon

Re: fd_sets building

On Tuesday 04 October 2005 19:39, Vincent Bruyere wrote:
> Good evening,
>
> I am working on an embedded MMS project with modem and
> PPP support.
> I am using lwip1.1.0 stack.
> I have several problems when receiving files bigger
> than 3 KBytes.
> The socket is often blocking to get data so i wonder
> if the socket select management is right.
>
> Does anybody know how to manage fd_sets building
> before running select, for a non UNIX OS ?
> ie is it necessary to use FD_ZERO, FD_SET, FD_ISSET
> like commands on real-time OS ?

lwIP sockets work just like UNIX sockets. That means that you should manage 
fd_sets just like you would on UNIX environment.

	- Atte
Vincent Bruyere | 5 Oct 2005 10:21
Picon
Favicon

Re: fd_sets building

ok Atte, thanks for your answer.
But as far as I know, the 'select' use is usefull on
the server side to avoid creating one process per
socket.
Is the 'select' routine necessary on the client side ?

Vincent

--- Atte Kojo <atte.kojo@...> a écrit :

> On Tuesday 04 October 2005 19:39, Vincent Bruyere
> wrote:
> > Good evening,
> >
> > I am working on an embedded MMS project with modem
> and
> > PPP support.
> > I am using lwip1.1.0 stack.
> > I have several problems when receiving files
> bigger
> > than 3 KBytes.
> > The socket is often blocking to get data so i
> wonder
> > if the socket select management is right.
> >
> > Does anybody know how to manage fd_sets building
> > before running select, for a non UNIX OS ?
> > ie is it necessary to use FD_ZERO, FD_SET,
> FD_ISSET
> > like commands on real-time OS ?
> 
> lwIP sockets work just like UNIX sockets. That means
> that you should manage 
> fd_sets just like you would on UNIX environment.
> 
> 	- Atte
> 
> 
> _______________________________________________
> lwip-users mailing list
> lwip-users@...
> http://lists.nongnu.org/mailman/listinfo/lwip-users
> 

	

	
		
___________________________________________________________________________ 
Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger 
Téléchargez cette version sur http://fr.messenger.yahoo.com
Durgaprasad Chivukula | 10 Oct 2005 11:34
Picon

[lwip users] using threads in our application safely ???

 
Hi Friends ,
 
 
         I already posted lot of questions to this group regarding my tftp client implementation of lwip , but I did not get any replies .I also posted my source code implementation , but the same. I hope you can guide me with my probelm this time . As I said I am writing a tftp client application on lwip using the socket API ,tapif interface (using bridge configuration ) on Redhat linux 2.4.20-8 . I need to put and get files from/ to a remote server. But both   put and get are not working.While getting I am only getting first 512 bytes and then tapif_input function halts I am creating only 1 thread in my application and I guess this is making problems , How can I use thread mechanism safely in my application ??? I am herewith attaching my implementation to this email , please have a look and suggest me in this ...Help!!!!!!!
 
 
 Also one more thing is I profile my whole stack using gprof /Functioncheck (profiling tools on linux ) as a part of my work , these tools are giving erroneous(wrong) results  for example sys_arch_sem_wait is called 744 times , sys_sem_signal is called 699 times ...and so on.... do u guys really think that these functions are called this many number of times.... I run my application and then run the tool , so it is a dynamic execution profiler.....I guess there is a problem with my thread implementation....
 
Please find below attached source code of my tftp client....in a zip file ... the important parts of the code to look at are
 
1. main_changed.c   (main function , main_thread function , setpeer , get ,put
2. tftp_changed.c    (tftp_sendfile , tftp_recvfile)
3.tftpsubs_changed.c (which implements the buffer  mechanism )
 
 
Please help me!!!!!
 
 
Regards
Durga
Attachment (TFTP_client.zip): application/zip, 21 KiB
_______________________________________________
lwip-users mailing list
lwip-users@...
http://lists.nongnu.org/mailman/listinfo/lwip-users
Christiaan Simons | 10 Oct 2005 13:10
Picon

Re: [lwip users] using threads in our application safely ???


lwip-users-bounces+christiaan.simons=axon.tv@... wrote
on 10-10-2005
11:34:25:

>          I already posted lot of questions to this group regarding
> my tftp client implementation of lwip , but I did not get any
> replies .I also posted my source code implementation , but the same.

Posting code doesn't make the coders actually look at it,
most of us dislike that. If you want implementation support
you should be more specific than 'it doesn't work' and describe
the problem in text please.

> I hope you can guide me with my probelm this time . As I said I am
> writing a tftp client application on lwip using the socket API ,
> tapif interface (using bridge configuration ) on Redhat linux 2.4.
> 20-8 . I need to put and get files from/ to a remote server. But
> both   put and get are not working.While getting I am only getting
> first 512 bytes and then tapif_input function halts

Ah this is better,
As far as I can remember the TFTP protocol uses a state-machine.
Can you tell us in which state the server is? And can you make
the state transitions visible?

Tapif_input will repeatedly be called for the incoming TFTP datagrams,
most likely causing server state changes.

> I am creating
> only 1 thread in my application and I guess this is making problems

Please do not guess, do you know whats actually going on in the code?

If there are multiple threads in the design you shouldn't remove
or disable them. What are these threads for?

> , How can I use thread mechanism safely in my application ??? I am
> herewith attaching my implementation to this email , please have a
> look and suggest me in this ...Help!!!!!!!
>
>
>  Also one more thing is I profile my whole stack using gprof
> /Functioncheck (profiling tools on linux ) as a part of my work ,
> these tools are giving erroneous(wrong) results  for example
> sys_arch_sem_wait is called 744 times , sys_sem_signal is called 699
> times ...and so on.... do u guys really think that these functions
> are called this many number of times.... I run my application and
> then run the tool , so it is a dynamic execution profiler.....I
> guess there is a problem with my thread implementation....

Ah, do not use gprof. For another project a colluege of mine
crippled some code by adding profiling flags to a Makefile.

It took me a week to find the root-cause.

Sorry, I'm not going to look at your code for now.

Christiaan Simons

Hardware / Software Engineer
Axon Digital Design

+31 (0)13 511 66 66
+31 (0)13 511 41 51

christiaan.simons@...
http://www.axon.tv

This email and any files transmitted with it are confidential and intended
solely for the use of the individual or entity to whom they are addressed.
If you have received this email in error please notify the system manager.
This message contains confidential information and is intended only for the
individual named.  If you are not the named addressee you should not
disseminate, distribute or copy this e-mail.
Durgaprasad Chivukula | 10 Oct 2005 14:52
Picon

Re: [lwip users] using threads in our application safely ???



On 10/10/05, Christiaan Simons <christiaan.simons-laxl2+jGqjQ@public.gmane.org> wrote:

lwip-users-bounces+christiaan.simons=axon.tv-qX2TKyscuCcdnm+yROfE0A@public.gmane.org wrote on 10-10-2005
11:34:25:

>Posting code doesn't make the coders actually look at it,
>most of us dislike that. If you want implementation support
>you should be more specific than 'it doesn't work' and describe
>the problem in text please.
 
      I mean to say that , I posted this question already , but did not get any reply .. I do not mean with the code...  

>Ah this is better,
>As far as I can remember the TFTP protocol uses a state-machine.
>Can you tell us in which state the server is? And can you make
>the state transitions visible?
 
 
     I use the TFTP server from the linux system , not from the lwip .While getting a file from the remote server,  I can observe the first 512 bytes (which tftp server has to send ) on my client side , using ethereal , but as I said  tapif_input halts after some time , I thought it would be the problem from my TFTP client timeout mechanism , so I increased the timeout of my TFTP client , no change.


>Please do not guess, do you know whats actually going on in the code?

>If there are multiple threads in the design you shouldn't remove
> or disable them. What are these threads for?
 
     I created only a single thread , there are no multiple threads , I saw in one of the lwip message list that we should create a new thread for sending and receiving , I do not know about my application , whether to use multiple threads . I have no idea of threading  , So I asked for an explanation of the thread mechanism which can be used safely in our applications,


>Ah, do not use gprof. For another project a colluege of mine
>crippled some code by adding profiling flags to a Makefile.
 
    Gprof , is ok with normal applications , but it has problems with threaded applications on Linux ,  Can we write an application on lwip  with out threading ...?`?? If so how can we do this...?
 

>It took me a week to find the root-cause.
 
   Where is the problem found , I mean the root cause of the probelm...???

>Sorry, I'm not going to look at your code for now.
 
  Please see the code when you find some time....
 
Regards
 
Durga  Prasad
Germany
 
 

 
 
     

Christiaan Simons

Hardware / Software Engineer
Axon Digital Design

+31 (0)13 511 66 66
+31 (0)13 511 41 51

christiaan.simons-laxl2+jGqjQ@public.gmane.org
http://www.axon.tv


This email and any files transmitted with it are confidential and intended
solely for the use of the individual or entity to whom they are addressed.
If you have received this email in error please notify the system manager.
This message contains confidential information and is intended only for the
individual named.  If you are not the named addressee you should not
disseminate, distribute or copy this e-mail.



_______________________________________________
lwip-users mailing list
lwip-users-qX2TKyscuCcdnm+yROfE0A@public.gmane.org
http://lists.nongnu.org/mailman/listinfo/lwip-users

_______________________________________________
lwip-users mailing list
lwip-users@...
http://lists.nongnu.org/mailman/listinfo/lwip-users
Christiaan Simons | 10 Oct 2005 16:35
Picon

Re: [lwip users] using threads in our application safely ???


>      I use the TFTP server from the linux system , not from the lwip
> .While getting a file from the remote server,  I can observe the
> first 512 bytes (which tftp server has to send ) on my client side ,
> using ethereal , but as I said  tapif_input halts after some time ,
> I thought it would be the problem from my TFTP client timeout
> mechanism , so I increased the timeout of my TFTP client , no change.

Ah, thus the TFTP client runs on top of lwip.

Does the server get a proper ACK from the client, in time?
The timeout of your server can play a major role here.

>      I created only a single thread , there are no multiple threads
> , I saw in one of the lwip message list that we should create a new
> thread for sending and receiving , I do not know about my
> application , whether to use multiple threads . I have no idea of
> threading  , So I asked for an explanation of the thread mechanism
> which can be used safely in our applications,

There are 2 examples in the contrib/ports/unix tree.
The 'simhost' example uses the netconn/socket api which you
probably use. Follow this as a reference.

>     Gprof , is ok with normal applications , but it has problems
> with threaded applications on Linux ,  Can we write an application
> on lwip  with out threading ...?`?? If so how can we do this...?

I just recommend not to use it.
Profiling won't help solving this issue anyway.

>   Please see the code when you find some time....
No.

Bye,

Christiaan Simons

Hardware / Software Engineer
Axon Digital Design

+31 (0)13 511 66 66
+31 (0)13 511 41 51

christiaan.simons@...
http://www.axon.tv

This email and any files transmitted with it are confidential and intended
solely for the use of the individual or entity to whom they are addressed.
If you have received this email in error please notify the system manager.
This message contains confidential information and is intended only for the
individual named.  If you are not the named addressee you should not
disseminate, distribute or copy this e-mail.

Gmane