On 7/16/19 11:28 AM, Holger Baust via 44Net wrote:
Hi.
My company plans to use DC/OS from Mesosphere and DC/OS is using 44.128.0.0/20 per default for overlay networking. This seems to have no impact since this overlay nets are only used in DC/OS clusters and not routed to the Internet, but might lead to problems in some cases...
Their documentation with this informations can be found at: https://docs.mesosphere.com/1.10/networking/virtual-networks/ip-per-containe...
You're posting using quoted printable, don't do that. I've fixed your link.
Is this already known?
News to me, and a really poor idea. There's other networks which could be used for this, without stepping on allocated space. 44.128.0.0/20 isn't used now, but neither was 1/8 or 5/8. I kinda want to announce 44.128.0.0/20 just to mess with them, might freak out if there's a route to that :)
On Tue, Jul 16, 2019 at 11:38:00AM -0400, Bryan Fields wrote:
On 7/16/19 11:28 AM, Holger Baust via 44Net wrote:
My company plans to use DC/OS from Mesosphere and DC/OS is using 44.128.0.0/20 per default for overlay networking. This seems to have no impact since this overlay nets are only used in DC/OS clusters and not routed to the Internet, but might lead to problems in some cases...
Their documentation with this informations can be found at: https://docs.mesosphere.com/1.10/networking/virtual-networks/ip-per-containe...
You're posting using quoted printable, don't do that. I've fixed your link.
Is this already known?
News to me, and a really poor idea. There's other networks which could be used for this, without stepping on allocated space. 44.128.0.0/20 isn't used now, but neither was 1/8 or 5/8. I kinda want to announce 44.128.0.0/20 just to mess with them, might freak out if there's a route to that :)
Yes, one should consider announcing 44.128.0.0/20. They are squatting on space which is not theirs to squat on.
In addition to announcing 44.128.0.0/20, it may be good to ensure all of 44.0.0.0/8 is visible in the default-free zone and drop any packets received for unused IP addresses within 44.0.0.0/8 on the floor.
The Internet default-free zone at times seems to adhere to a "use it or lose it" mantra, and while this is unfortunate; announcing the 44net to the internet at large will send a clear message that the space is actually in use and should not be squatted on.
To completely cover the 44net block from a BGP perspective, in addition to the existing 44.0.0.0/9 and 44.128.0.0/10 announcement, 44.192.0.0/10 should also be announced (and nullrouted) from some vantage point.
I'm happy to do so if the community accepts my help in this regard. I have facilities to announce the prefix and ignore any traffic received for it.
Kind regards,
Job Snijders
On Jul 16, 2019, at 11:00 AM, Job Snijders job@ntt.net wrote:
To completely cover the 44net block from a BGP perspective, in addition to the existing 44.0.0.0/9 and 44.128.0.0/10 announcement, 44.192.0.0/10 should also be announced (and nullrouted) from some vantage point.
I'm happy to do so if the community accepts my help in this regard. I have facilities to announce the prefix and ignore any traffic received for it.
If this is decided to be a good idea, I can help too - I also have facilities and and submit the routes as a customer route to 3356 and 174 (and 2914 but I expect Job has this well covered <g>) which might help from a “customer routes” vs. “peer routes” perspective.
Adam K0RAB BGP Enthusiast ;-)
I can also help in this endeavor if required, but I think the /20 should be a bit split into two parts (2 x /21) for example, just to be more specific (in getting the point across). :-)
On 16 Jul 2019, at 19:11, Adam Korab ak@mid.net wrote:
On Jul 16, 2019, at 11:00 AM, Job Snijders job@ntt.net wrote:
To completely cover the 44net block from a BGP perspective, in addition to the existing 44.0.0.0/9 and 44.128.0.0/10 announcement, 44.192.0.0/10 should also be announced (and nullrouted) from some vantage point.
I'm happy to do so if the community accepts my help in this regard. I have facilities to announce the prefix and ignore any traffic received for it.
If this is decided to be a good idea, I can help too - I also have facilities and and submit the routes as a customer route to 3356 and 174 (and 2914 but I expect Job has this well covered <g>) which might help from a “customer routes” vs. “peer routes” perspective.
Adam K0RAB BGP Enthusiast ;-)
44Net mailing list 44Net@mailman.ampr.org https://mailman.ampr.org/mailman/listinfo/44net
On Tue, Jul 16, 2019 at 04:11:30PM +0000, Adam Korab wrote:
If this is decided to be a good idea, I can help too - I also have facilities and and submit the routes as a customer route to 3356 and 174 (and 2914 but I expect Job has this well covered <g>) which might help from a “customer routes” vs. “peer routes” perspective.
Adam (and others),
Please don't advertise 44.192.0.0/10. To do so would cause some fairly significant problems. I am preparing a document to explain what's up with this and hope to have it ready in a week or so. I ask your patience. - Brian
44/8 is by default routed to UCSD, so the /20 part IS already routed. Announcing the /20 would be useless imo
73,
Ruben - ON3RVH
-----Original Message----- From: 44Net 44net-bounces+on3rvh=on3rvh.be@mailman.ampr.org On Behalf Of Brian Kantor Sent: dinsdag 16 juli 2019 18:20 To: AMPRNet working group 44net@mailman.ampr.org Subject: Re: [44net] Questions regarding usage off 44/8 parts outside of ampr.org
On Tue, Jul 16, 2019 at 04:11:30PM +0000, Adam Korab wrote:
If this is decided to be a good idea, I can help too - I also have facilities and and submit the routes as a customer route to 3356 and 174 (and 2914 but I expect Job has this well covered <g>) which might help from a “customer routes” vs. “peer routes” perspective.
Adam (and others),
Please don't advertise 44.192.0.0/10. To do so would cause some fairly significant problems. I am preparing a document to explain what's up with this and hope to have it ready in a week or so. I ask your patience. - Brian
_________________________________________ 44Net mailing list 44Net@mailman.ampr.org https://mailman.ampr.org/mailman/listinfo/44net
-
Adam Korab -
Antonios Chariton -
Brian Kantor -
Bryan Fields -
Job Snijders -
Ruben ON3RVH