[Community-Discuss] Let's chat about unannounced v6 space!

[Andrew Alston]

Hi Nishal,

Not as far as I know, but I could be wrong – I don’t think I’ve seen a single page output of stuff not announced and who it was allocated to – could be wrong though.

As regards another full view – I could use another full view that is in IOS-XR format (from a show route ipv6) and accessible via ssh – the current code connects to the router and does an ssh2_exec with no PTY allocation to pull the table. (save me writing more code to read other formats and change the methods – so if you got anything like that it would be useful! – Otherwise if you can send me off list some output from whatever routers you have that have full table that aren’t XR when I have time I can take a look at how to match against formatting)

If you can send me the list of IXP prefixes that shouldn’t be visible that would be awesome – I can write a whitelist fairly easily, and yes, its potentially possible for me to flag visible that shouldn’t be.

As a note – the v4 processing code is almost done – I was playing with this last night and I can already see partial announcements against the delegation file – the next step is tying that back to org-id’s and organisations and then publishing in similar format.  There are a coupla caveats behind that in the fact that the delegation file in AfriNIC for some strange reason is not on natural CIDR boundaries and the whois database also doesn’t store CIDR masks so some rather painful processing and aggregation has to happen before you actually do the comparison, but there should be a similar page for v4 out in the next few days.

If you have a list of v4 IXP prefix’s as a result, this would be useful as well.

Chat later

Andrew

On 17/10/2016, 19:09, "Nishal Goburdhan" <nishal at controlfreak.co.za> wrote:

    On 16 Oct 2016, at 10:56, Andrew Alston wrote:
    
    > Just as another note –
    >
    > I am also working on something that will take the rest of the 
    > allocated prefixes (the ones that are announced), and tie them back to 
    > the originating ASN
    
    
    hi andrew,
    
    doesn’t afrinic do this already (they used to, with the AIRRS report, 
    but i haven’t looked in a long time).
    
    that said, it’s always nice to see initiative, and perhaps you can do 
    something new, so, by way of assistance, i offer:
    * since you’re doing this from one “full-view”, do you need 
    another?
    * i can assist you quite easily with identifying the IXP peering 
    prefixes, if you’d like  (i.e those that should not be visible.  for 
    bonus points, perhaps you can flag those, that should *not* be, but are 
    ;-))
    
    —n.