Deduplication on replay attacks allows messages from a different concentrator to pass

[spygi]

Consider the scenario:

• Unconfirmed message with frame counter 10, Deduplication Mark or None
• Station A picks it up, LNS marks it as Not Duplicate, goes upstream ✔
• Station B gets it, LNS sees that it was already handled, marks it as SoftDuplicate, goes upstream ✔
• Station A gets it again because of replay. LNS marks it as Resubmission, does not go upstream because it's marked as invalid counter in ValidateRequest ✔
• Station B gets it again because of replay. LNS marks it as SoftDuplicate, else if (payloadFcnt == loRaDevice.FCntUp && concentratorDeduplicationResult == ConcentratorDeduplicationResult.SoftDuplicateDueToDeduplicationStrategy) is true this time, goes upstream. ❌

There could be potentially other similar scenarios that are not covered besides this one.

Acceptance criteria:

• Ensure test coverage for replay attacks. As of now we only check if message is sent only 1 additional time (total of 2).

[p-schuler]

@spygi - can you validate this again?