Skip to Content.
Sympa Menu

rare-dev - Re: [rare-dev] polka is approaching rare... was: Re: custom crc32 library in freerouter

Subject: Rare project developers

List archive

Re: [rare-dev] polka is approaching rare... was: Re: custom crc32 library in freerouter


Chronological Thread 
  • From: mc36 <>
  • To: Cristina Klippel Dominicini <>
  • Cc: Rafael Silva Guimarães <>, "" <>, "Moises R. N. Ribeiro" <>, Magnos Martinello <>
  • Subject: Re: [rare-dev] polka is approaching rare... was: Re: custom crc32 library in freerouter
  • Date: Wed, 27 Oct 2021 22:42:37 +0200

hi,
sorry for the spam... so i introduced polka in freerouter with this huge
change:
https://github.com/mc36/freeRouter/commit/2f0981910316236abc044eaaffa694a531bb7ff3
unfortunately for now, the forwarding can only work with the cryPoly class,
which is a nicer version i previously sent to you... the issue i'm facing
right now is that i cannot figure out how to apply crc16 properly...
i'm generating the S values and compute a routeid that i can decode with the S
according to your paper, but when it comes to applying crc16, i fail
somehow...
the value that come out after the final xor is invalid...
i've written a small tester for that at http://src.nop.hu/src/zzz.java
it generates 10 16bit polynomials starting at 32768, those are the Sn
values...
then encodes the path 1->3 3->5 5->7 7->8 8->9... then it decode with the
poly class, those are the On values... whose appear on the left side,
those right side seems correct... the rest have garbage, obviously...
but when it come to crc16, it fails to decode correctly...:(
any idea is highly welcome! :)
thanks,
cs

here is how you can try it yourself:

$ wget src.nop.hu/rtr.zip > /dev/null
$ unzip rtr.zip
$ cd src
src$ ./c.sh
src$ ./zzz.sh

it gives you the following:

s0=1000000000000000
s1=1000000000000001
s2=1000000000000011
s3=1000000000000101
s4=1000000000001001
s5=1000000000010001
s6=1000000000010111
s7=1000000000011011
s8=1000000000100001
s9=1000000000101101
o0=0x04a6
o1=0x0003
o2=0x7371
o3=0x0005
o4=0x5da0
o5=0x0007
o6=0x2ce9
o7=0x0008
o8=0x0009
o9=0x0000
r= 00 00 00 00 00 e7 52 d4 f9 b9 4a b7 a9 0a 04 a6
crc=2254 dif=04a6 xor=26f2




On 10/27/21 14:28, mc36 wrote:
hi,
as said before, i came up with my own poly.java which is specific to gf(2),
had add,sub,mul,div and modinv...
at the moment it have an S generator and a checker... below is a test run,
first some basic poly math, then
your paper's calculation, which is fine, then a randomly generated S and O,
encoding of R and verification of it...
everything seems fine at least as i see, but i have issues with your
polka.patch... there, you propose the numbers
0x2b,0x2d,0x39,0x3f,0x47,0x53,0x8d,0xbd,0xd7,0xf5, i checked some and it
seems to me that these are not irreducible...
for an example, i picked up randomly 0x3f and 0x53 and according to my poly
lib, it have a gcm of 1001... you can verify it at
http://www.ee.unb.ca/cgi-bin/tervo/calc.pl?num=111111&den=1001&f=d and
http://www.ee.unb.ca/cgi-bin/tervo/calc.pl?num=1010011&den=1001&f=d
both of them give back 000 as remainder...
as for now, i have everything, i'll slowly integrate it to freerouter to see
how it fits... :)
seeking for your comments,
regards,
cs





[..]




Archive powered by MHonArc 2.6.19.

Top of Page