I've been doing a lot of reading and practicing with IPSec and DMVPN setups lately. One thing that's routinely been a point of confusion for me has been ISAKMP rekeying policies. It seems simple enough, when DH is used, a secure management connection exists to share protected cryptographic key information for the standup or continuance of protected communications. I.e., when that management connection times out, no more key material can be shared until a new management connection is in place.
That's a strong assumption. Assume for discussion we're using an old static crypto map setup, with PSK and pure static entries in tunnel mode. If traffic arrives to your border that doesn't match your ACL exactly, traffic will still be allowed across the wire, but it'll be unencrypted.
Let's say I setup an isakmp policy with a lifetime of 200 seconds and in my crypto map I set a security-association policy with a lifetime of 125 seconds. When my ipsec keys approach their end of lifetime, the isakmp management connection will be used to rekey both sides of the connection, correct? What happens when that isakmp policy expires? In my mind, a new connection should be negotiated to provide for continuance of the ipsec rekeying activities. If that were to fail, the keys shouldn't be sent. Or would they be?
I've tested this exhaustively in my labs and to be honest, I'm getting none the wiser. I setup the exact scenario I laid out above: isakmp 200 and ipsec 125 second lifetimes, respectively. I can see the isakmp informational messages coming across when the isakmp policies expire and eventually I will see a new connection when I issue a 'show crypto isakmp sa', but it's not immediate. In fact, on some iterations, the isakmp sa showed me a (deleted) status, yet as I watch the ipsec SAs with 'show crypto ipsec sa' I see each respective SPI's lifetime reach its end-of-life and then as if a gremlin is inside the router, boom, rekeyed.
I've gone through all the RFCs I can stand to read, Deal's entire VPN book, multiple cisco publications, mountains of forums, and I can't understand this simple scenario. It seems logical that phase 2 cannot happen without phase 1, yet I sit here and watch these SAs rekey without the isakmp SA showing me it's been refreshed.
Maybe I've confused the smallest detail somewhere. I'm hoping someone can help me fan this fog out a little.
I would lean toward agreeing with you, however, everything I can find about isakmp lists it as the necessary part of IPSec needed to trigger phase 2. To me, that would indicate that phasse one is required to provide the initial keying and subsequent keying.
Login to the FXOS chassis manager.
Direct your browser to https://hostname/, and log-in using the user-name and password.
Go to Help > About and check the current version:
Check the current version availa...
We have configured the outside and inside Interface with official ipv6 adresses, set a default route on outside Interface to our router, we also have definied a rule , which also gets hits, to permit tcp from inside Interface to any6.
In Syslog I also se...