ipv6 in companies…
ipv6 is not hard, but for internal networking no company (really) “needs” more than rfc1918 address space.
thus any decision in that direction is always “less” needed than any bonus for (da)magement personnel is crucial for the whole companies survival…
for companies services to be reachable from outside/ipv6 mostly “only” the loadbalancers/revproxies etc need to be ipv6 ready but … this i.e. also produces logs that possibly break decades old regexes that no one understands any more (as the good engineers left due to too many boni payed to damagement personnel) while other access/deny rules that could break or worse let through where they should block (remember that 192.168. could the local part of ipv6 IF sone genious used a matching mech that treats the dot “.” as a wildcard as overpayed damagement personnel made them rush too fast), could be hidden “somewhere”. altogether technical debt is a huge blocker for everything, especially company growth, and if no customer “demands” ipv6, then it stays on the damagement personnels list as “fulfilling the whishes of engineers to keep them happy” instead of on the always deleted “cleaning up technical debt caused by damagement personnel” list.
setting up firewalls for ipv6 is quite easy and if you go the finegrained “whitelisted or drop/block” approach from the beginning it might take a bit for ipv6 specials to be known to you, but the much bigger thing is IMHO the then current state of firewall rules. and who knows every existing rule? what rules should be removed already and must not be ported to ipv6? usually firewalls and their rules are a big mess due to … again too many boni payed to damagement personnel, hindering the company from the needed steps forward…
ipv6 adoption is slow for reasons that are driving huge cars that in turn speed up other problems ;-|
i once had to look at a firefall appliance cluster, (discovered, it could not do any failover in its current state but somehow the decider was ok with that) but when looking at its logs, i discovered an rsh and rcp access from an ip address that belonged to a military organisation from a different continent. i had to make it a security incident. later the vendor said that this was only the cluster internal routing (over the dedicated crosslink), used for synchronisation (the thing that did not work) and was only used by a separate routing table only for clustersync and that could never be used for real traffic.
but why not simply use an ip that you “own” by yourself and PTR it with a hint about what this ip is used for? instead of customers scratching their head why military still uses rcp and rsh. i guess because no company reads firewall logs anyway XD
someone elses ip? yes! becuase they’ll never find out !!1!
i really appreciate that ipv6 has things like a dedicated documentation address range and that fc00:/7 is nicely short.
ipv6 in companies… ipv6 is not hard, but for internal networking no company (really) “needs” more than rfc1918 address space. thus any decision in that direction is always “less” needed than any bonus for (da)magement personnel is crucial for the whole companies survival…
for companies services to be reachable from outside/ipv6 mostly “only” the loadbalancers/revproxies etc need to be ipv6 ready but … this i.e. also produces logs that possibly break decades old regexes that no one understands any more (as the good engineers left due to too many boni payed to damagement personnel) while other access/deny rules that could break or worse let through where they should block (remember that 192.168. could the local part of ipv6 IF sone genious used a matching mech that treats the dot “.” as a wildcard as overpayed damagement personnel made them rush too fast), could be hidden “somewhere”. altogether technical debt is a huge blocker for everything, especially company growth, and if no customer “demands” ipv6, then it stays on the damagement personnels list as “fulfilling the whishes of engineers to keep them happy” instead of on the always deleted “cleaning up technical debt caused by damagement personnel” list.
setting up firewalls for ipv6 is quite easy and if you go the finegrained “whitelisted or drop/block” approach from the beginning it might take a bit for ipv6 specials to be known to you, but the much bigger thing is IMHO the then current state of firewall rules. and who knows every existing rule? what rules should be removed already and must not be ported to ipv6? usually firewalls and their rules are a big mess due to … again too many boni payed to damagement personnel, hindering the company from the needed steps forward…
ipv6 adoption is slow for reasons that are driving huge cars that in turn speed up other problems ;-|
deleted by creator
i once had to look at a firefall appliance cluster, (discovered, it could not do any failover in its current state but somehow the decider was ok with that) but when looking at its logs, i discovered an rsh and rcp access from an ip address that belonged to a military organisation from a different continent. i had to make it a security incident. later the vendor said that this was only the cluster internal routing (over the dedicated crosslink), used for synchronisation (the thing that did not work) and was only used by a separate routing table only for clustersync and that could never be used for real traffic. but why not simply use an ip that you “own” by yourself and PTR it with a hint about what this ip is used for? instead of customers scratching their head why military still uses rcp and rsh. i guess because no company reads firewall logs anyway XD
someone elses ip? yes! becuase they’ll never find out !!1!
i really appreciate that ipv6 has things like a dedicated documentation address range and that fc00:/7 is nicely short.