networking - tcp reno, newreno and slow start -

When a loss is packet loss at the beginning, does the Reno / Nereano algorithm possibly report duplicates, or is it Purely slow -> RTO?

In this way, if two packets are sent (at the beginning of a slow start), and one disappears first, then what starts at slow but RTO?

This is misleading, because RFC says that in practice, they apply (slow start and rescue from the crowd) simultaneously 'and the Linux source is a bit rough and only one implementation.

When the packet loss occurs at a slower pace initially, what is probably the duplex notices Reno / Are near the neurone algorithm, or does it start at a slow speed -> RTE?

I will say "yes", duplicate ACK will be detected and action will be taken.

In this way, if two packets are sent (at the beginning of a slow start), and one disappears first, then what starts at slow but RTO?

This simple example will be "simple RTO" during the beginning of a slow start when only two packets can be sent, then one duplicate ACK (incoming) The trigger by the second packet will be the most). Both the packets will be accepted simultaneously, so there can also be anyone. But a duplicate ACK does not react quickly. Therefore TCP will wait for re-marriage timer to end the deadline.

This is misleading, because RFC says that in practice, they apply (slow start and save from congestion) simultaneously 'and the Linux source is slightly read and only one implementation .

I agree that the Linux source is a rough read. But this is definite and if you need to know actually then its only option may be :) Until you read (or wrote) nobody will find it; Which is not me