On Wed, 29 August 2001, Patrick Greenwell wrote:
The difference is the carrier's response, in particular how well they keep their customers informed.
That is certainly *a* point of differentiation, however if the goal of these "basement dual-homers" is to not suffer downtime due to the outage of a single provider(much like the organizations that "matter"), all the responsiveness in the world from a provider whose circuit to one of the "basement dual-homers" which has failed isn't going to prevent them from being down, is it?
I don't know what the goal of "basement dual-homers" is. I think that is the other thread, which I haven't been keeping up with. It is possible to reduce the impact of most provider problems. I had a seven year streak without a service affecting outage at my NOC longer than the BGP route convergence interval. Even "hot-standby, fail-over" systems require a finite time to recover. Even with all the network insanity between 1993 and 2000, I was always able to reach the working portion of the Internet. If my own connection didn't work, how could I post about those problems. It was a lot of hard work, and I'll admit I had a very steep learning curve. And I'll be the first to admit, there were force maejure events my network wouldn't be able to handle. But those were my choices, and I could make those decisions with accurate information. The first step of operating a reliable network is choosing carriers willing to keep you, their customer, fully informed. Only after you have accurate information from the carriers can you make informed decisions concerning your network. Most of my biggest problems occurred due to inaccurate or lack of information from a carrier. If I know about the risk, I could choose to mitigate it or choose to do nothing. If the carrier doesn't keep me informed about jeopardy conditions in their network, managing that risk is more difficult. A side-effect of the lack of information is the extrodinary measures customers feel they must take to protect their network connectivity. Do you really need to multi-home, have your own AS number, have your own CIDR block for reliable network connectivity? With DHCP, you could just unplug from one provider and plug into the new one, and auto-magically have connectivity. With Dynamic DNS update you could automagically redirect your DNS name to your new address. IP addresses could be assigned geographically, instead of by provider, and aggregation could take place anywhere from the census tract level to the hemisphere level.
On 30 Aug 2001, Sean Donelan wrote:
On Wed, 29 August 2001, Patrick Greenwell wrote:
The difference is the carrier's response, in particular how well they keep their customers informed.
That is certainly *a* point of differentiation, however if the goal of these "basement dual-homers" is to not suffer downtime due to the outage of a single provider(much like the organizations that "matter"), all the responsiveness in the world from a provider whose circuit to one of the "basement dual-homers" which has failed isn't going to prevent them from being down, is it?
I don't know what the goal of "basement dual-homers" is. I think that is the other thread, which I haven't been keeping up with.
My point was simply that the "basement dual-homers" probably want the same thing as the "people that matter", which is disparate paths to the Internet. *snipped a lot of things I agree with*
With DHCP, you could just unplug from one provider and plug into the new one, and auto-magically have connectivity. With Dynamic DNS update you could automagically redirect your DNS name to your new address. IP addresses could be assigned geographically, instead of by provider, and aggregation could take place anywhere from the census tract level to the hemisphere level.
Except for the small problem that there is no guarantee that DNS servers will honor TTLs, that this solution would necessitate dialing down TTLs to the lowest possible value to maximize effectiveness which would greatly increase DNS-related traffic, and flapping could be really ugly. DNS in lieu of a routing protocol make me uncomfortable.
participants (2)
-
Patrick Greenwell
-
Sean Donelan