I have updated the RIB estimation model to include growth rates 
for IPv4 and IPv6 individually.
Link for an updated set of slides based on this refinement is:
http://www.antd.nist.gov/~ksriram/BGPSEC_RIB_Estimation_v4_v6.pdf 

Tony has suggested 60% or higher growth rate for IPv6 prefixes.
Currently, I am assuming 100%, 90%, 80%, etc. in the initial years
with gradual decline to 30% and then keep it steady at 30% yearly growth.
This amounts to a cumulative growth of nearly 25000% 
from 2011 to 2025 for eBGP IPv6 prefixes.

For IPv4, I am using a steady 12% yearly growth rate, which is
slightly more aggressive than what is suggested by 
Geoff Huston's measurement data over the 2008-2011 period.

The above assumptions can be revised based on updated measurements
as they become available (including any effects of IPv4 deaggregation 
due to IPv4 exhaustion).
 
Chris mentioned that operators and vendors should play with the
Excel spreadsheet model to have conversations about planning for the
RIB requirements. I am happy to share the spreadsheet. 
Here is the link for it: 
http://www.antd.nist.gov/~ksriram/BGPSEC_RIB.xls  

The spreadsheet is very amenable to inputting your view of the
IPv4, IPv6 growth rates and BGPSEC adoption curve,
and you can see the resulting RIB estimates for route reflectors and PE routers.
You can also observe the RIB size comparisons between 
RSA-2048 and ECDSA-256 signature algorithms 
for your various growth assumptions.

Thanks for the discussion and comments so far. 
Further comments/suggestions are very welcome.

Sriram
NIST  

_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr

Reply via email to