Hey,

Going through a design effort on a new very large enterprise/provider hi-bred network.

When reviewing various documents such as 'High Availability Campus Network Design-Routed Access Layer...' and 'IP Multicast Best Practices for Enterprise Customers' along with various CCIE study guides and other non-Cisco docs, the OSPF process 'timers throttle spf' recommended defaults are 10,100-500,5000 msec.

The only references I find with something different are 'Service Module Design with ACE and FWSM' and 'Cisco Service Delivery Center Infrastructure 2.1 Design Guide'. These two references might be considered more closely aligned with Service Provider environments. In neither of these documents is there any discussion as to why the 1000 1000 1000 recommendation is being made. They just list recommended configs for the test-beds with these values.

Question: why are the 1000 1000 1000 settings be suggested?

It's hard for me to imagine a scenario where you want to wait a full second to initiate a change, then force OSPF to only and ever use 1 second for a min/max for consecutive recalcs. There must be some practical reason why 1000 1000 1000 makes sense in a provider/data center environment when 10,100,5000 is being recommended for the enterprise. And what about a very large enterprise with data centers?

Any comments or opinions welcome.