Nick> Abstract:
Nick> The world doesn't run on the stuff you read about on hacker news. It runs 
on
Nick> solaris boxes hidden in the heart of infrastructures, on mainframes 
almost old
Nick> enough for social security, and on metal in racks. If it has endured it 
was
Nick> well built and has had years to be refined and perfected. The business 
needs
Nick> haven't changed, nor has the logic behind this stuff, but increasingly the
Nick> skills required to build, expand, tend to and maintain it are scarce. 
Security
Nick> becomes a matter of exploits being forgotten, not patched. Interfaces are 
dated
Nick> and slow, bottlenecking things and killing agility. Eventually, for these
Nick> reasons and others you may find yourself just looking straight at the 
systems
Nick> version of the colliseum and going - damn, what I really need there is a
Nick> skyscraper.

Nick> So how do you do it? How do you go from unmanaged, artisinal bespoke 
machines
Nick> to config managed? How do you go from configuration managed 
infrastructure to
Nick> more dynamic? From there to modern patterns? The urban renewal reference 
wasn't
Nick> an accident - how do you avoid doing this in a way that parallels the 
story of
Nick> places like Boston's own West End? Oh, and how the hell do you bolt on 
tests
Nick> when your service count makes most people's node count look small and 
you're
Nick> fairly sure your dependency chain requires non-euclidean math and sorting 
it
Nick> might unleash eldritch horrors?

Nick> I may or may not have answers to these, but I've battled this particular 
tire
Nick> fire a few times before.

I'd love to attend, but I've got prior commitments for Scouts.  Can
someone video this talk, or will there be notes, etc to be shared?
John

_______________________________________________
bblisa mailing list
[email protected]
http://www.bblisa.org/mailman/listinfo/bblisa

Reply via email to