Re: [sidr] WGLC for draft-ietf-sidr-rpki-rtr-rfc6810-bis-03

[David Mandelberg <[e-mail]>]

On 2015-03-17 14:29, George, Wes wrote:
> This
> may be as simple as recommending that in the case where data from 
> multiple
> caches is held and specific entries conflict with one another, there
> SHOULD be an odd number of caches so that there is basis for 
> comparison to
> determine which cache is out of sync or providing incorrect info. 
> (i.e.
> Have 3 so that you can go with the 2/3 that agree)

Are you suggesting comparison of all the data from each single cache as 
an atomic entity, or comparison of individual IPvX and Router Key PDUs?

If the former, then I think that would work fine as long as a majority 
(or maybe even a plurality) of the caches has the exact same data. But 
what does the router do if this is not the case? If the caches all 
download from the RPKI at different times, then I would expect it to be 
common for no two caches to have the same data.

If the latter, then the semantics depend heavily on exactly how the 
comparison is done. Lets say a CA simultaneously issues one ROA for {AS 
65536, 10.0.0.0/8} and another for {AS 65537, 10.0.0.0/8}. Some of the 
caches see the publication point before both ROAs are issued; some see 
the pub point after both ROAs are issued and published. Can you 
guarantee that the voting mechanism will always result in either both 
ROA payloads, or neither, being used? (If a router ends up using one but 
not the other, then a previously unknown route becomes invalid.)

-- 
David Eric Mandelberg / dseomn
http://david.mandelberg.org/