On Mon, 20 Jun 2005, Joe Abley wrote:
On 2005-06-20, at 18:49, jamie.baddeley(a)vpc.co.nz wrote:
Some good points Michael, but I suspect your analysis doesn't include failure doesn't include a visit from Mr FatFingers and friend at the Operations console.
This is a wildly good point. Vijay Gill did an analysis of AOL's Nx10G backbone network which he presented at a NANOG a while back, and it transpired that the overwhelming majority of customer- affecting outages were due to operator error, rather than equipment or circuit failure.
I seem to remember somebody (I completely forget who) touting one advantage of their new network being that the core configuration was so simple and straightforward that it would (almost) never need to be changed. Thus nobody would ever login and potentially break it [1]. Possibly the fact that networking *does* involve actual physical equipment plugging into physical ports makes it harder to automate and drive everything from a database like other areas. [1] - On the other hand when you do have the problem and it takes longer to fix since people aren't used to working with the system. See the letter in this week's computerworld. -- Simon J. Lyall. | Very Busy | Mail: simon(a)darkmere.gen.nz "To stay awake all night adds a day to your life" - Stilgar | eMT.