Rate of growth on IPv6 not fast enough?

Bryan Fields Bryan at bryanfields.net
Mon Apr 19 17:51:48 UTC 2010


On 4/19/2010 13:40, David Conrad wrote:
> Bryan,
> 
> On Apr 19, 2010, at 10:22 AM, Bryan Fields wrote:
>> Here is some unverified calculations I did on the problem of scaling nat.
>>
>> Right now I'm using 42 translation entries in my nat table.  Each entry takes
>> up 312 bytes of FIB memory, which is ~12.7 Kib of data in the FIB.  Mutiply
>> this by 250k users and we have 3,124,237 KiB of FIB entries, or 3.1 GiB.  This
>> is not running any PtP programs or really hitting the network, I'm just
>> browsing the web and typing this email to you.
> 
> This is really interesting data.  What hardware is this on?

Cisco.  I've not had an engineer look at it, but it's based on this FAQ:
http://www.cisco.com/en/US/tech/tk648/tk361/technologies_q_and_a_item09186a00800e523b.shtml

Q. How many concurrent NAT sessions are supported in Cisco IOS NAT?

A. The NAT session limit is bounded by the amount of available DRAM in the
router. Each NAT translation consumes about 312 bytes in DRAM. As a result,
10,000 translations (more than would generally be handled on a single router)
consume about 3 MB. Therefore, typical routing hardware has more than enough
memory to support thousands of NAT translations.

Anyone from the vendors want to speak up and maybe poke some holes in my math?
I'd actually love to be wrong about the amount of memory for this, but suspect
I'm close :(

-- 
Bryan Fields

727-409-1194 - Voice
727-214-2508 - Fax
http://bryanfields.net




More information about the NANOG mailing list