TestGround tests for higher usefulness interval

[aarshkshah1992]

In #651:

We introduced a change to double the usefulness interval i.e. the grace period we give to peers to be deemed "useful"(faster response to queries) before we replace them in the Routing Table in favour of a new peer.

Earlier, an attacker could create too many connections to us causing us to disconnect peers we have in our RT. This would then affect their "usefulness" scores(since we measure usefulness by query response latency) as the query time would have to pay the connection cost for them. Hence, we'd evict them through no fault of theirs, thus giving the attacker space in the RT.

Now, that's harder because you'd have to cause more disconnects to cause a low usefulness score for a a prolonged duration.

However, this could hurt performance as we'll keep slow peers in the RT around for longer too.

We could test all this using the TG. Need to discuss if it's worth it though.

[Stebalien]

We've gotten rid of usefulness so this is no longer relevant.