On 6/11/20, 1:19 PM, "Saku Ytti" <[email protected]> wrote:
On Thu, 11 Jun 2020 at 19:49, Phil Bedard <[email protected]> wrote:
> As for normal v6 forwarding, the way most higher speed routers made
recently work there is little difference in latency since the encapsulation for
the packet is done in a common function at the end of the pipeline and the
lookups are often in the same memory space. NPUs are also being built today
with enough on-package memory to hold larger routing tables. Whether a packet
has to be buffered on-chip vs. off-chip has a much larger impact on latency/PDV
than a forwarding lookup.
On-package is not important, on-chip or off-chip is what matters, i.e.
do you eat SERDES to connect memory or not.
[pmb]
Sorry meant to say on-die, not on-package.
Typically the time it takes to do those lookups are built into the system specs
to attain the performance you need with deterministic latency within a certain
bounds. There are certainly corner cases where you make tradeoffs, especially
now that single NPUs are 10+ Tbps, but it's not really an MPLS vs. IPv4 vs.
IPv6 thing. The other key is to do those types of accesses in a single pass,
not traverse multiple hierarchy levels or do multiple operations. If you are
tunneling then the table for any of those types is going to small on a
mid-point router.
