Charles Wu wrote:
Can you please explain how this is applicable in modern-day implementations
of TCP? From my limited understanding, Nagle is a relic of the past (been
replaced by TCP Westwood, etc)
Nagle is very old circa 1984 I believe, but it hasn't really be
replaced. Many folks would choose to use other algorithms for queuing in
high throughput links, but generally nagle is on by default. Clearly,
some form of queuing is desirable for maximum throughput of small
packets, but more interactive applications are hurt by queuing e.g.
VoIP. Therefore, it is useful to see what throughput is obtained with
and without the setting on if you are considering using the radio pair
for VoIP.
Yes, the bit is turned on, but can you please explain how this is applicable
for a transparent layer-2 bridging scenario?
It isn't applicable for a layer-2 bridging scenario. However, it can
affect layer-3 devices on either side of the bridge when doing the
throughput test, which may have an impact on the test. In my experience,
it does not.
-Matt
--
WISPA Wireless List: [email protected]
Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless
Archives: http://lists.wispa.org/pipermail/wireless/
