Rare project developers

[mc36 <>]

hi,
i go inline...
regards,
cs


On 10/28/21 03:05, Cristina Klippel Dominicini wrote:
> Hi Csaba,
>
> That is amazing progress!!!
thanks! it was just your good paper and good sample code that helped a lot!
some more progess here, as you're carrying ethertype within polka header, i 
added
some mpls over polka test cases...
so one can tunnel mpls vpns (both layer2 and layer3) through a polka domain...
https://github.com/mc36/freeRouter/commit/aa261c3e3a5bf33f85e8068cf7555653f24c4e85


> I shared with you our repository with the preliminary code translation to 
> Java, which comprises the routeid computation for a given path (represented 
> by a set of nodeids and its respective portids). Today I taught classes all 
> day, and was not able to provide appropriate feedback, but tomorrow I will 
> translate the part related to the generation of arbitrary irreducible 
> polynomials for the nodeids and the CRC-based forwarding.
sounds interesting, especially the array of irreducible polys that i'm 
interested at the moment...

> > > but interestingly, i had to comment out the irreducible check: throw new 
> > > Exception("NodeID polynomial is not irreducible");
> If you are getting this exception, it means that your nodeID polynomials are 
> not irreducible, which is a requisite for using the Chinese Remainder Theorem.
at the moment, i pick the n+1 entry and check if it have all gcm <= 1 to thet 
previous n entries...
you can find in previous mails the S array for your polkaMainTest.java, imho 
those polys are irreducible... :)
but all in all, you know math better than me, sooo i'm open to change my 
mind/code... :))

> Regarding the design decisions (tables, MAC rewrite, indexes...), I agree with your previous considerations. I will elaborate some discussion points for tomorrow. 
sounds interesting, waiting for your upcoming mails...

> The discussion in the VC yesterday with you, Frederic and Jordi highlighted 
> very interesting arguments to exploit PolKA's benefits in real-world 
> deployments.
thanks... just for your information, i enabled polka in my homenet 
networkwide and things seems normal...
i can pass packets over real interfaces.... at the moment they're completely 
handled in java's crc16 code,
but i'll progress soon with the p4dpdk dataplane export to see how if works 
in emulated hardware...
(that one will enable some more test cases concentrating on polka dataplane, 
so after that, i can
progress merging your patch to the tofino code and bring that to all tests 
passing level...)

sid#show config-differences
router bgp4 65535
 no redistribute connected
 exit
interface tunnel1
 no description
 tunnel vrf v1
] tunnel source loopback0
 tunnel destination 10.10.10.1
 tunnel domain-name 10.10.10.5 10.10.10.20 10.10.10.199
 tunnel mode polka
 vrf forwarding v1
 ipv4 address 1.1.1.2 255.255.255.252
 no shutdown
 no log-link-change
 exit

sid#ping 1.1.1.1
pinging 1.1.1.1, src=10.10.10.227, vrf=v1, cnt=5, len=64, tim=1000, gap=0, 
ttl=255, tos=0, flow=0, fill=0, sweep=false, multi=false, detail=false
!!!!!
result=100%, recv/sent/lost/err=5/5/0/0, rtt min/avg/max/total=30/31/33/159
sid#show interfaces tun1
tunnel1 is up (since 00:00:31, 2 changes)
 description:
 type is polka, hwaddr=none, mtu=1500, bw=8000kbps, vrf=v1
 ip4 address=1.1.1.2/30, netmask=255.255.255.252, ifcid=619101103
 received 0 packets (0 bytes) dropped 2 packets (132 bytes)
 transmitted 15 packets (990 bytes) promisc=false macsec=false

sid#debug iface polka
sid#ping 1.1.1.1
pinging 1.1.1.1, src=10.10.10.227, vrf=v1, cnt=5, len=64, tim=1000, gap=0, 
ttl=255, tos=0, flow=0, fill=0, sweep=false, multi=false, detail=false
debug ifcPolka.send2eth:ifcPolka.java:263 tx ttl=255 proto=2048 route= 00 00 
00 00 00 00 00 00 22 54 dc cf 40 10 9e 59
!debug ifcPolka.send2eth:ifcPolka.java:263 tx ttl=255 proto=2048 route= 00 00 
00 00 00 00 00 00 22 54 dc cf 40 10 9e 59
!debug ifcPolka.send2eth:ifcPolka.java:263 tx ttl=255 proto=2048 route= 00 00 
00 00 00 00 00 00 22 54 dc cf 40 10 9e 59
!debug ifcPolka.send2eth:ifcPolka.java:263 tx ttl=255 proto=2048 route= 00 00 
00 00 00 00 00 00 22 54 dc cf 40 10 9e 59
!debug ifcPolka.send2eth:ifcPolka.java:263 tx ttl=255 proto=2048 route= 00 00 
00 00 00 00 00 00 22 54 dc cf 40 10 9e 59
!
result=100%, recv/sent/lost/err=5/5/0/0, rtt min/avg/max/total=31/32/35/164
sid#





> Best regards,
> Cristina
> ________________________________________
> De: mc36 <>
> Enviado: quarta-feira, 27 de outubro de 2021 20:23
> Para: Cristina Klippel Dominicini
> Cc: Rafael Silva Guimarães; ; Moises R. N. Ribeiro; 
> Magnos Martinello
> Assunto: Re: polka is approaching rare... was: Re: custom crc32 library in 
> freerouter
>
> hi,
> sorry for the spam... finally managed to find out what was the issue with my 
> crc issue:
> the degree of the S polynominals i used was 15 instead of 16... now i changed 
> a bit the
> configs to start the polys at 64k and now thing seems fine here with crc16 
> computation...
> https://github.com/mc36/freeRouter/commit/fe7f83f7464fe4383c733795d2919eacb8e3e98f
> i'll do some more tests but if things are normal then i'll proceed tomorrow 
> with the
> much simpler task: to export the iface-->coeff and the 
> (neighboring)nodeid-->nexthop
> tables from freerouter... then the next step will be to introduce polka to 
> the p4dpdk
> dataplane... i run a 30 node homenet on top of that dataplane so it'll be the 
> next
> level of testing here, in live network... :)
> regards,
> cs
>
> mc36@safe:/data.pub/src$ ./zzz.sh
> s0=10000000000000000 17
> s1=10000000000000001 17
> s2=10000000000000011 17
> s3=10000000000000101 17
> s4=10000000000001001 17
> s5=10000000000001111 17
> s6=10000000000010001 17
> s7=10000000000011011 17
> s8=10000000000011101 17
> s9=10000000000101011 17
> o0=0xfbe8
> o1=0x0003
> o2=0x544e
> o3=0x0005
> o4=0x989f
> o5=0x0007
> o6=0x6cd7
> o7=0x0008
> o8=0x0009
> o9=0x0000
> r= 00 00 00 00 be 43 e1 8c ed 0f 43 ed 0a c6 fb e8
> i=0 crc=0000 dif=fbe8 xor=fbe8
> i=1 crc=fbeb dif=fbe8 xor=0003
> i=2 crc=afa6 dif=fbe8 xor=544e
> i=3 crc=fbed dif=fbe8 xor=0005
> i=4 crc=6377 dif=fbe8 xor=989f
> i=5 crc=fbef dif=fbe8 xor=0007
> i=6 crc=973f dif=fbe8 xor=6cd7
> i=7 crc=fbe0 dif=fbe8 xor=0008
> i=8 crc=fbe1 dif=fbe8 xor=0009
> i=9 crc=fbe8 dif=fbe8 xor=0000
>
>
> On 10/28/21 00:38, mc36 wrote:
> > hi,
> > thanks for adding me... i've cross-checked the calculators and found that 
> > yours and mine create the same routeid...
> > did some cosmetical changes, the printer functions and the sX and oX arrays 
> > have been updated in your code...
> > below are the runs from yours and mine, both calculate the same routeid: 
> > 0x999e998f5578001
> > but interestingly, i had to comment out the irreducible check: throw new 
> > Exception("NodeID polynomial is not irreducible");
> > egards,
> > cs
> >
> >   Â Â Â Â Â Â Â  long s1Array[] = {1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
> > 0, 0, 0, 0};
> >   Â Â Â Â Â Â Â  long s2Array[] = {1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
> > 0, 0, 0, 1};
> >   Â Â Â Â Â Â Â  long s3Array[] = {1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
> > 0, 0, 1, 1};
> >   Â Â Â Â Â Â Â  long s4Array[] = {1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
> > 0, 1, 0, 1};
> >
> >   Â Â Â Â Â Â Â  long o1Array[] = {0, 0, 0, 1};
> >   Â Â Â Â Â Â Â  long o2Array[] = {0, 0, 1, 0};
> >   Â Â Â Â Â Â Â  long o3Array[] = {0, 0, 1, 1};
> >   Â Â Â Â Â Â Â  long o4Array[] = {0, 0, 0, 0};
> >
> >
> >
> > Oct 28, 2021 12:19:52 AM polka.PolKA <init>
> > FINE: PolKA Constructor
> > Oct 28, 2021 12:19:52 AM polka.PolKA <init>
> > FINE: S = [x^15, 1+x^15, 1+x+x^15, 1+x^2+x^15]
> > Oct 28, 2021 12:19:52 AM polka.PolKA <init>
> > FINE: O = [1, x, 1+x, 0]
> > Oct 28, 2021 12:19:52 AM polka.PolKA calculateR
> > INFO: PolKA RouteID Computation
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: S0 = 1000000000000000 = 0x8000
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: S1 = 1000000000000001 = 0x8001
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: S2 = 1000000000000011 = 0x8003
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: S3 = 1000000000000101 = 0x8005
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: O0 = 0001 = 0x1
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: O1 = 0010 = 0x2
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: O2 = 0011 = 0x3
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: O3 = 0000 = 0x0
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: t0 = 1000000000000111000000000001001000000000001111 = 0x2001c004800f
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: t1 = 1000000000000110000000000001111000000000000000 = 0x200180078000
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: t2 = 1000000000000100000000000000101000000000000000 = 0x200100028000
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: t3 = 1000000000000010000000000000011000000000000000 = 0x200080018000
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: n0 = 11001100110011 = 0x3333
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: n1 = 1000000000000 = 0x1000
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: n2 = 100101010101011 = 0x4aab
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: n3 = 110100110011010 = 0x699a
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: xx0 = 11001100110010001100110011010010101010101001000000000000001 = 
> > 0x666466695548001
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: xx1 = 10000000000001100000000000011110000000000000000000000000000 = 
> > 0x4003000f0000000
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: xx2 = 1101111111111011111111111110011010000000001001000000000000000 = 
> > 0x1bff7ffcd0048000
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: xx3 = 0000 = 0x0
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: x = 1100110011001000010011001101010110101010100000000000000000001 = 
> > 0x1999099ab5500001
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: m = 1000000000000111000000000001001000000000001111000000000000000 = 
> > 0x1000e00240078000
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: f = 0100110011001111010011001100011110101010101111000000000000001 = 
> > 0x999e998f5578001
> > Oct 28, 2021 12:19:52 AM polka.PolKA calculateR
> > FINE: Check[0]:
> > Oct 28, 2021 12:19:52 AM polka.PolKA calculateR
> > FINE: == true
> > Oct 28, 2021 12:19:52 AM polka.PolKA calculateR
> > FINE: Check[1]:
> > Oct 28, 2021 12:19:52 AM polka.PolKA calculateR
> > FINE: == true
> > Oct 28, 2021 12:19:52 AM polka.PolKA calculateR
> > FINE: Check[2]:
> > Oct 28, 2021 12:19:52 AM polka.PolKA calculateR
> > FINE: == true
> > Oct 28, 2021 12:19:52 AM polka.PolKA calculateR
> > FINE: Check[3]:
> > Oct 28, 2021 12:19:52 AM polka.PolKA calculateR
> > FINE: == true
> > Oct 28, 2021 12:19:52 AM polka.PolKA printPoly
> > INFO: R = 0100110011001111010011001100011110101010101111000000000000001 = 
> > 0x999e998f5578001
> > BUILD SUCCESSFUL (total time: 0 seconds)
> >
> >
> >
> >
> > s0=1000000000000000
> > s1=1000000000000001
> > s2=1000000000000011
> > s3=1000000000000101
> > s4=1000000000001001
> > s5=1000000000010001
> > s6=1000000000010111
> > s7=1000000000011011
> > s8=1000000000100001
> > s9=1000000000101101
> > t0 = 2001c004800f
> > n0 = 3333
> > t1 = 200180078000
> > n1 = 1000
> > t2 = 200100028000
> > n2 = 4aab
> > t3 = 200080018000
> > n3 = 699a
> > xx0 = 666466695548001
> > xx1 = 4003000f0000000
> > xx2 = 1bff7ffcd0048000
> > xx3 = 0
> > x = 1999099ab5500001
> > o0=0x0001
> > o1=0x0002
> > o2=0x0003
> > o3=0x0000
> > o4=0x0f26
> > o5=0x27db
> > o6=0x2df6
> > o7=0x3ca4
> > o8=0x4382
> > o9=0x1d24
> > r= 00 00 00 00 00 00 00 00 09 99 e9 98 f5 57 80 01
> > crc=d581 dif=8001 xor=5580
> > mc36@safe:/data.pub/src$
> >
> >
> >
> >
> >
> >
> > On 10/27/21 23:25, mc36 wrote:
> > > hi,
> > > ohh, i forgot to mention that if you're interested to execute the tests then 
> > > do the following:
> > >
> > > $ wget src.nop.hu/rtr.zip
> > > $ unzip rtr.zip
> > > $ cd src
> > > src$ ./c.sh
> > > src$ java -jar rtr.jar test tester rout-bgp529 wait
> > >
> > > it'll start the processes for the bgp based polka test...
> > > then it'll pause at the end... then in some new windows, you can access the 
> > > routers:
> > >
> > > $ telnet localhost 20001Â Â Â Â  #for r1
> > > $ telnet localhost 20002Â Â Â Â  #for r2
> > > $ telnet localhost 20003Â Â Â Â  #for r3
> > >
> > > not too much to see there, just bare minimal at the moment:
> > > - show ipv4 srindex v1Â Â Â Â Â Â  #to see the nodeid to prefix mapping 
> > > table... in the dataplanes, it'll be a shortcut to a nexthop...
> > > - show ipv6 srindex v1Â Â Â Â Â Â  #to see the nodeid to prefix mapping 
> > > table... in the dataplanes, it'll be a shortcut to a nexthop...
> > > - display inter eth1 ether   #here you can see the polka packets 
> > > received/sent counters per interface... ctrl+q to stop viewing...
> > > - debug iface polka          #to activate polka packet 
> > > debugging on a given router...
> > > - term mon                   #to see the 
> > > debugs interactive
> > > - show logg                  #to see the 
> > > historical debugs
> > >
> > > regards,
> > > cs
> > >
> > >
> > > On 10/27/21 22:42, mc36 wrote:
> > > > 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
> > > >
> > > > [..]
>
>
> ________________________________
>
> Esta mensagem (incluindo anexos) contém informação confidencial destinada 
> a um usuário específico e seu conteúdo é protegido por lei. Se você não 
> é o destinatário correto deve apagar esta mensagem.
>
> O emitente desta mensagem é responsável por seu conteúdo e endereçamento.
> Cabe ao destinatário cuidar quanto ao tratamento adequado. A divulgação, 
> reprodução e/ou distribuição sem a devida autorização ou qualquer outra 
> ação sem conformidade com as normas internas do Ifes são proibidas e 
> passíveis de sanção disciplinar, cível e criminal.