: if someone cannot get out somewhere, they're obviously
: going to get in touch with me as to why. Once this is
: done, it is explained
: I've always contacted someone
: after about 3 attempts at getting someone to assess
: their network
I know from experience this doesn't scale into the hundreds of thousands of customers and can only imagine the big ass eyeball network's scalability issues...
scott
: if someone cannot get out somewhere, they're obviously
: going to get in touch with me as to why. Once this is
: done, it is explained: I've always contacted someone
: after about 3 attempts at getting someone to assess
: their networkI know from experience this doesn't scale into the hundreds of thousands of customers and can only imagine the big ass eyeball network's scalability issues...
scott
Hear hear...
Scaling process and procedures is often as hard or harder than scaling technical things...
Unfortunately, the lesson that scaling either is hard is only really something that one can learn through experience -- I know that I for one used to believe (as I would bet did most of us) that you could scale just by buying a bigger X, where X could be a router, circuit, etc. If that didn't work you could always just buy another X (or a bunch more Xs) -- this strategy works up to a point, after which it all goes pear-shaped. Until you have experienced this firsthand it is hard to truly understand.
The same thing happens with things like abuse -- it is easy to deal with abuse on a small scale. It is somewhat harder on a medium scale and harder still on a large scale -- the progression from small to medium to large is close to linear. At some point though the difficulty suddenly hockey-sticks and becomes distinctly non-trivial -- this doesn't mean that it is impossible, nor that you should give up, but rather that a different approach is needed. Understanding this is harder than understanding why you cannot grow your network just by buying more X.
W
First, I don't buy this. I think dealing with abuse is *much*
easier for large operations than small.
But suppose you're right. Let me concede that point for the purpose
of making my second point (and generic "you" throughout, BTW):
Second, I don't really care how hard it is. It's YOUR network, YOU
built it, YOU plugged it into our Internet: therefore, however hard
it is, it's YOUR problem. Fix it.
Or if you choose not to: at least stop whining about how much you
don't like the way in which other people try to partially compensate
for YOUR failure.
---Rsk
> I know from experience this doesn't scale into the hundreds of
> thousands of customers and can only imagine the big ass eyeball
> network's scalability issues...
Hear hear...
Scaling process and procedures is often as hard or harder than
scaling technical things...
It's true. But the big networks hire people who understand scaling
issues and know how to make things work. It's not up to us to solve
their scaling problem. If you can define a mechanism that will work on
smaller networks to achieve a goal, and if that goal is worthwhile
achieving, the the big networks will get their scalability networks to
scale it up. There is a similar problem in chemicals where researchers
create new compounds in the laboratory and then hand the details over to
scaling experts who know how to change the process to work on the scale
of a factory. And it's not unusual to see chemical factories that are
acres in size.
The same thing happens with things like abuse -- it is easy to deal
with abuse on a small scale. It is somewhat harder on a medium scale
and harder still on a large scale -- the progression from small to
medium to large is close to linear. At some point though the
difficulty suddenly hockey-sticks and becomes distinctly non-trivial
-- this doesn't mean that it is impossible, nor that you should give
up, but rather that a different approach is needed. Understanding
this is harder than understanding why you cannot grow your network
just by buying more X.
Yes this is true. But the people who find different approaches need to
see how the smaller networks solve a problem. Their skill is not in
finding solutions to abuse, but in figuring out how to restructure an
abuse solution to work on a huge scale.
--Michael Dillon