TP6.1 - LRC: Enforcing unresponsive Application Server blacklist policy.

TP6.1 - LRC: Enforcing unresponsive Application Server blacklist policy.

Important notice.

This notice is a reminder of the TP6 release note document.

If you have an Application Server (in ThingPark Wireless Device Manager) that uses “BLAST” routing strategy with multiple destinations, you could be impacted.

  • From now on, “BLAST” routing strategy is effectively implemented as an in-sequence multicast.
    • each destination (i.e. each external AS url) is processed, one by one.
    • should an application server become unresponsive or slow enough, it would trigger a blacklist mechanism, therefore rendering the entire set of destinations not usable for a while.
      • this mechanism is inherited from SEQUENTIAL routing strategy

→ if you still need an in-parallel “BLAST”-like behaviour, we suggest that you create as many single-destination Application Servers as necessary.

{
BLAST AS1:
  - url_01
  - url_02
  - ...
  - url_n
}

should become instead:

SEQUENTIAL AS_01:
  - url_01

SEQUENTIAL AS_02:
  - url_02

[...]

SEQUENTIAL AS_n:
  - url_n

  • hint: you know you are actually impacted if your various application servers do not receive the same amount of traffic, relative to one another.
  • for more details, please refer to RDTP-5790 section of ThingPark Wireless 6.0 release note document.