Re: [Karlnet] Which F/W?

[Chris Conn <cconn@abacom.com>]

Caleb Carroll wrote:
> Could be.  I would suggest you downgrade to 4.31 firmware and see if the problem still exists.  4.43k is whack.

Hello,

4.43K may cause some problems for some people (I don't recall any being 
specifically described in this list...), however I as well as a number of ISPs 
find it to solve many problems regarding radios going "deaf" as Karlnet coined 
it.  We have upgraded the totality of our network to 4.43K and we have no 
problems at all.  We can focus on many other things since this version is 
available from Karlnet, and I can positively say that it corrects instability 
problems we have been suffering from for months.

As well, the polling algorithm of 4.43K is much superior to that of 4.31, and 
anyone using a multipoint base station system with a number of satellites will 
see a very large throughput gain and overall system responsiveness gain.

The only finding I have with it is that it loads tertiary firmware of 8.74 to 
the Orinoco radios.  We have found this to cause mild to severe retransmit 
percentages in the base stations.  The simple fact of upgrading to 4.43K (and 
therefore the radio card to 8.74) made all satellites jump from 0-0.5% 
retransmits to 2.5-5% retransmits across the board.  Anyone else notice that 
their retransmit counters are higher with 4.43K?

I did some tests with some 4.x builds I have that don't load any tertiary 
firmware, and the retransmit problem does not exist when the base station radio 
is using 8.10, however 8.72 and 8.74 firmware seems prone to a retransmit 
condition of some sort.

Chris