Re: Shim6 failure recovery after garbage collection

["Per Heldal" <heldal@eml.cc>]

Igor,

Throughout your message you refer to overhead/session. Shim6 operates at
L3, so overhead must be counted per unique source-dest host (source-dest
ULID-pair). Neither does shim6 have to re-negotiate or change state for
each protocol-connection that gets established or torn down. Not saying
there is no overhead, but this makes a huge difference to the numbers.

Shim6 will, if or when it becomes available, represent an opportunity to
equipment-makers like any other new technology. That will normally lead
to the development of new products and product-families. It's premature
to dismiss a technology during research just because something is done
differently.

The biggest mistake in recent months is that shim6 has been used as an
argument in discussions of operational nature (RIR policies). Devices
generally don't run very well on vapor-ware, and thus isn't very popular
among operators ;-)  OTOH, operational concerns and feedback from the
ops-community is important to shim6 development.


//per

-- 
  Per Heldal
  http://heldal.eml.cc/