number of hops != performance

Does anyone have a nice reference I can point to to once and for all


that just because a customer has 6-8 L3 hops within our network (all at
gigabit speeds or higher) that doesnt automatically mean they are getting
bad performance or higher latency.

When I was at Ebone I would sometimes visit large accounts just to help
the salespeople with this question. I would bring along a copy of our
measurement data and show these customers exactly what our network latency
was measured in milliseconds not in hops. Everyone understood milliseconds
and I knew that our network had lower latency than the competition. Of
course, the fact that we did measure latency with a matrix of over 20
cities was a strong selling point as was the fact that we made this data
available to customers. The salespeople were overjoyed because this was
far stronger than the material that our marketing department made

Hiding the L3 hops in a MPLS core (or other L2 switching) doesnt mean
customers are getting better performance since equipment today forwards
just as quickly on L3 as on L2.

And this is exactly what I would explain. Customers are intelligent enough
to understand that hops equal devices and that hiding the hops/devices
from IP traceroutes does not make them go away. Then I told them that we
engineered our network for speed (wirespeed switching, zero buffering,
zero packet loss) so that we weren't ashamed of the hops. But perhaps the
other guys had something to hide?

When you have something great to sell, you have to cut through the
bullshit and demonstrate the greatness to the customer. Assuming that they
want low latency, you have to show them beyond a shadow of a doubt that
your latency is low. And if your latency isn't low, then you need to
either fix the problem or find customers who care about other things like
PoP locations or SLA reporting or something other than network quality.

--Michael Dillon