Chris,
I appreciate your viewpoint, but we have seen examples where ping times
grow into SECONDS, and that simply is NOT acceptable.
Other than that, I do agree with your point!
Kevin
**************************************
Kevin R. Knuth
Business Development Manager
North America
260-424-9690 Regional Office
614-822-5275 Corporate Office
kknuth@karlnet.com
www.karlnet.com
**************************************
-----Original Message-----
From: karlnet-bounces@WISPNotes.com
[mailto:karlnet-bounces@WISPNotes.com] On Behalf Of Chris Conn
Sent: Thursday, September 11, 2003 4:26 PM
To: Karlnet Mailing List
Subject: Re: [Karlnet] RE: Ping Jitter
> Also, there seems to be some debate about whether this
> is a problem or not - can you clarify if Karlnet see
> this as a problem or just the nature of how polling
> works?
Hello,
I won't speak for Karlnet, however it is mostly a cosmetic problem in
the minds
of the people who decide that a few 200ms RTT = slow. However, there is
some
discussion as to low SNR showing poor throughput, which I have not seen
in my
networks. I have not personally seen this, probably because I like to
set the
satellites at 5.5Mbps and this increases their sensitivity. Have the
15dB or
less people tried this? You can still leave the base at 11Mbps. I even
have
some very poor (< 10db) satellites at 2Mbps; this sacrifices some
upstream
speed, however they can download at the full rate with fewer problems
(ie fewer
retransmits from the base).
Karlnet is consistently working to improve their polling algorithm, and
if they
can make it "look" faster to the people who like to see consistent 30ms
or less
on IDLE satellites without sacrificing possible throughput gains on
active
satellites, good for them. I personally have not seen the need (other
than
the educational one) to stabilize rtt when the client is doing nothing!
The fact of the matter is, ICMP round trip times are not a benchmark of
anything other than the fact the packet made it all the way back. They
can be
used as a guide or indication of latency, but you also have to
understand the
transport media (in this case, Turbocell) to judge if a few stray 300ms
packets
compared to relative traffic levels truly indicate poor performance.
Again, believe what you want and throw away your Karlnet and buy other
products
if that is your inclination. You can also study your network with
better tools
and make up your mind as to the throughput you can obtain from it.
Chris
_______________________________________________
Karlnet mailing list
Karlnet@WISPNotes.com
http://lists.wispnotes.com/mailman/listinfo/karlnet
|