What to do about the network I
What problem are we trying to solve?
1) We have broadcasts going around and polluting the ATP->Bufferbox connections.
ATP transfers compete with those broadcasts. We need to trim them down to a affordable level. It’s all HV. Next Workshop. The transfers do NOT compete with other traffic - they are segregated by switches, remember.
$ time /usr/sbin/tcpdump -i eth0 -c 500 dst 130.199.99.255 and not src phoncs0
0.19user 1.35system 12:36.22elapsed 0%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (217major+180minor)pagefaults 0swaps
So without the HV, you have a 0.66 Hz broadcast rate.
287 out of the 500 broadcasts came from the ATPs and SEB’s (netbios)
$ time /usr/sbin/tcpdump -i eth0 -c 500 dst 130.199.99.255 and src phoncs0
0.01user 0.04system 0:11.84elapsed 0%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (215major+85minor)pagefaults 0swaps