X HB should not be disabled until after confirmations
X New algo for slowing down reduction of sat stuff.
X prohibit sat stuff if ever the RTT goes under 500ms
X Should a HB be sent to a both dest on t3? - No just lower threshold
X bind should zero out when doing ifa_ifawithaddr() so
  that if the user forgets it will still find the ifa.
- random numbers for ephemerial ports.
+ need to make sctplib a .so as well as the .a
- need to be sure no cwnd advance is done with any
  data chunk marked for retran. Since its the old
  dest (which you don't know) that actually deserves
  the bump.
- Alternate route code should be written so that
  we go in and attempt to diversify the routes at
  the SCTP level after the load_addresses_from_init()
  is run. This will allow us to have as diverse routes
  as possible.