«

Hi,

Does any one know a resource which provides some statistics about IPv4
and IPv6 addresses stored in a forwarding table of a router.
For Example,
for IPv4, each entry in a forwarding table contains at least the
destination address and the corresponding next hop. Destination
addresses might be stored entirely or they can match a set of IP
addresses. What I am looking for is some statistics that can tell me
how on the average how many entries are of the form
127.123.65.42
136.121.45.12x (where 'x' means a don't care i.e. use the longest
prefix matching here)
136.121.45.1xx (where 'x' means a don't care i.e. use the longest
prefix matching here)
136.141.54.xxx (where 'x' means a don't care i.e. use the longest
prefix matching here)
136.141.5x.xxx (where 'x' means a don't care i.e. use the longest
prefix matching here)
and so on.

I will really appreciate any help.

Thanks.

In article <. com>,
gullu <> wrote:

>Does any one know a resource which provides some statistics about IPv4
>and IPv6 addresses stored in a forwarding table of a router.
>For Example,
>for IPv4, each entry in a forwarding table contains at least the
>destination address and the corresponding next hop. Destination
>addresses might be stored entirely or they can match a set of IP
>addresses. What I am looking for is some statistics that can tell me
>how on the average how many entries are of the form
>127.123.65.42
>136.121.45.12x (where 'x' means a don't care i.e. use the longest
>prefix matching here)
>136.121.45.1xx (where 'x' means a don't care i.e. use the longest
>prefix matching here)
>136.141.54.xxx (where 'x' means a don't care i.e. use the longest
>prefix matching here)
>136.141.5x.xxx (where 'x' means a don't care i.e. use the longest
>prefix matching here)
> and so on.

Your examples hint that you might be misunderstanding the routing
table. "longest prefix matching" does not use the decimal address
values: it uses the binary addresses, the "longest" is based upon
the number of bits matched, not the number of decimal digits matched.

Cisco routers do not keep statistics about the routing table size
(and on Cisco routers, the routing process can end up being -much-
more complex than a simple forwarding table.) However, you can
generate these kinds of statistics yourself by using snmp to
periodically read out the forwarding table and doing some simple
totalling. You want the IP-FORWARD-MIB, base OID 1.3.6.1.2.1.4.24
In particular, you might want ipCidrRouteTable 1.3.6.1.2.1.4.24.4
or the inetCidrRouteTable 1.3.6.1.2.1.4.24.7
http://tools.cisco.com/Support/SNMP/...6.1.2.1.4.24.7

You may wish to investigate the IFM IPv6 / IPv4 web page.

You've heard a lot of talk about IPv6, and you'd like to have a play
with it, but your ISP doesn't offer it. This article discusses one
option, 6to4, that you can use.

http://www.ifm.net.nz/research/ipv6.html

as well as Microsoft SBS 2003 IPv6 Readiness:

http://www.ifm.net.nz/research/sbs20...readiness.html

and the Cisco IPv6 Config Wizard:

http://www.ifm.net.nz/cookbooks/ipv6configwizard.html

Sincerely,

Brad Reese
Cisco Resumes
http://www.bradreese.com/cisco-resumes.htm

You may also wish to view:

http://sourceforge.net/search/?type_...oft&words=IPv4

Sincerely,

Brad Reese
http://www.BradReese.Com

Hi Walter,

Thanks for your reply.

I know that the prefix matching is done on binary addresses. I just
gave an example so that everyone can understand it.

I was wondering if CISCO collected some data about the distribution of
the IP addresses in the forwarding table on order to introduce some
optimizations in their future products.

Can you answer one more question? Do CISCO routers use TCAMs for IP
lookup in the forward table. If so (and if you feel that you are
allowed to answer this question), Can you provide me with some details
about how these TCAMs are used in the routers. There are so many
research papers out on IP lookup and TCAMs and I am just wondering what
industry has to say about it.

Thanks.

On Jan 26, 11:45 am, rober...@hushmail.com (Walter Roberson) wrote:
> In article <1169830758.286385.202...@q2g2000cwa.googlegroups. com>,
>
>
>
> gullu <atifhas...@gmail.com> wrote:
> >Does any one know a resource which provides some statistics about IPv4
> >and IPv6 addresses stored in a forwarding table of a router.
> >For Example,
> >for IPv4, each entry in a forwarding table contains at least the
> >destination address and the corresponding next hop. Destination
> >addresses might be stored entirely or they can match a set of IP
> >addresses. What I am looking for is some statistics that can tell me
> >how on the average how many entries are of the form
> >127.123.65.42
> >136.121.45.12x (where 'x' means a don't care i.e. use the longest
> >prefix matching here)
> >136.121.45.1xx (where 'x' means a don't care i.e. use the longest
> >prefix matching here)
> >136.141.54.xxx (where 'x' means a don't care i.e. use the longest
> >prefix matching here)
> >136.141.5x.xxx (where 'x' means a don't care i.e. use the longest
> >prefix matching here)
> > and so on.Your examples hint that you might be misunderstanding the routing
> table. "longest prefix matching" does not use the decimal address
> values: it uses the binary addresses, the "longest" is based upon
> the number of bits matched, not the number of decimal digits matched.
>
> Cisco routers do not keep statistics about the routing table size
> (and on Cisco routers, the routing process can end up being -much-
> more complex than a simple forwarding table.) However, you can
> generate these kinds of statistics yourself by using snmp to
> periodically read out the forwarding table and doing some simple
> totalling. You want the IP-FORWARD-MIB, base OID 1.3.6.1.2.1.4.24
> In particular, you might want ipCidrRouteTable 1.3.6.1.2.1.4.24.4
> or the inetCidrRouteTable 1.3.6.1.2.1.4.24.7http://tools.cisco.com/Support/SNMP/do/BrowseOID.do?local=en&translat...

"gullu" <> wrote in message
news: oups.com...
> Hi Walter,
>
> Thanks for your reply.
>
> I know that the prefix matching is done on binary addresses. I just
> gave an example so that everyone can understand it.
>
> I was wondering if CISCO collected some data about the distribution of
> the IP addresses in the forwarding table on order to introduce some
> optimizations in their future products.

it wouldnt make sense - there is no way my routers are sending forwarding
table info to Cisco....

more importantly - why would that help?

Cisco routers get used in all sorts of environments, many of which arent
even directly connected to the Internet
>
> Can you answer one more question? Do CISCO routers use TCAMs for IP
> lookup in the forward table. If so (and if you feel that you are
> allowed to answer this question), Can you provide me with some details
> about how these TCAMs are used in the routers. There are so many
> research papers out on IP lookup and TCAMs and I am just wondering what
> industry has to say about it.

depends. low end boxes use software (but may use TCAM table structures).

Cisco have put TCAM logic into chips for higher performance boxes such as
the 7609 with a high end processor card.

the search for "tcam forwarding table" on the cisco web site gets 1680
hits - without a CCO login....
>
> Thanks.
>
> On Jan 26, 11:45 am, rober...@hushmail.com (Walter Roberson) wrote:
> > In article <1169830758.286385.202...@q2g2000cwa.googlegroups. com>,
> >
> >
> >
> > gullu <atifhas...@gmail.com> wrote:
> > >Does any one know a resource which provides some statistics about IPv4
> > >and IPv6 addresses stored in a forwarding table of a router.
> > >For Example,
> > >for IPv4, each entry in a forwarding table contains at least the
> > >destination address and the corresponding next hop. Destination
> > >addresses might be stored entirely or they can match a set of IP
> > >addresses. What I am looking for is some statistics that can tell me
> > >how on the average how many entries are of the form
> > >127.123.65.42
> > >136.121.45.12x (where 'x' means a don't care i.e. use the longest
> > >prefix matching here)
> > >136.121.45.1xx (where 'x' means a don't care i.e. use the longest
> > >prefix matching here)
> > >136.141.54.xxx (where 'x' means a don't care i.e. use the longest
> > >prefix matching here)
> > >136.141.5x.xxx (where 'x' means a don't care i.e. use the longest
> > >prefix matching here)
> > > and so on.Your examples hint that you might be misunderstanding the
routing
> > table. "longest prefix matching" does not use the decimal address
> > values: it uses the binary addresses, the "longest" is based upon
> > the number of bits matched, not the number of decimal digits matched.
> >
> > Cisco routers do not keep statistics about the routing table size
> > (and on Cisco routers, the routing process can end up being -much-
> > more complex than a simple forwarding table.) However, you can
> > generate these kinds of statistics yourself by using snmp to
> > periodically read out the forwarding table and doing some simple
> > totalling. You want the IP-FORWARD-MIB, base OID 1.3.6.1.2.1.4.24
> > In particular, you might want ipCidrRouteTable 1.3.6.1.2.1.4.24.4
> > or the inetCidrRouteTable
1.3.6.1.2.1.4.24.7http://tools.cisco.com/Support/SNMP/do/BrowseOID.do?local=
en&translat...
>
--
Regards

- replace xyz with ntl

In article < .com>,
gullu <> wrote:

>Can you answer one more question? Do CISCO routers use TCAMs for IP
>lookup in the forward table. If so (and if you feel that you are
>allowed to answer this question), Can you provide me with some details
>about how these TCAMs are used in the routers.

Any relevant NDA I had expired long ago, so I can answer with
absolute confidence: "I haven't the foggiest idea."