Glosario de red
Los términos que aparecen al leer un traceroute, explicados con claridad.
- Traceroute
- A traceroute maps the path a packet takes to reach a destination. It sends packets with a deliberately short lifetime, so each router along the way is forced to answer, revealing itself one step at a time.
- Hop / saut
- One router on the path. Hop 1 is usually your own gateway; the last hop is the destination. A hop that does not answer shows as a timeout, which is normal — many routers are configured not to reply.
- TTL
- Time To Live: a counter carried by every packet. Each router decrements it by one, and when it reaches zero the router discards the packet and reports back. Traceroute abuses this on purpose to discover routers one by one.
- Latency / latence
- The round-trip time between you and a hop, in milliseconds. It grows with physical distance — roughly 1 ms per 100 km of fibre — so a sudden jump usually means the traffic crossed an ocean or a congested link.
- Jitter / gigue
- The variation in latency between successive measurements. Steady 80 ms is fine for a video call; 30 ms that swings to 200 ms is not. Jitter, not average latency, is what makes calls choppy and games feel unfair.
- Packet loss
- The share of probes a hop never answered. Loss at an intermediate hop alone often means nothing — routers deprioritise replies to themselves. Loss that persists from one hop through to the destination is the real signal.
- ASN
- Autonomous System Number: the identifier of a network operated by a single organisation. Seeing the ASN of each hop tells you whose network your traffic is crossing, and where it changes hands.
- BGP
- Border Gateway Protocol: how networks announce to each other which addresses they can reach. It decides the route your packets take between operators, and a bad announcement elsewhere in the world can reroute your traffic without warning.
- MTR
- A tool that combines traceroute and ping: instead of one pass, it keeps probing every hop and accumulates statistics. That is what makes intermittent loss visible where a single traceroute would show a clean path.
- ICMP
- The protocol routers use for control messages, including the replies traceroute depends on. Some networks rate-limit or block it, which is why a UDP or TCP traceroute sometimes gets further than an ICMP one.
- Reverse DNS
- Reverse DNS turns an IP address back into a name. Operators often encode the city and link type in it, so a hostname like ae-3.par-th2.example.net quietly tells you the hop is in Paris.
- CGNAT
- Carrier-Grade NAT: when an operator shares one public address between many subscribers. It explains hops in the 100.64.0.0/10 range near the start of a trace, and it complicates anything that needs an inbound connection.
- Peering
- A direct interconnection between two networks so their traffic does not have to transit a third. Good peering is what keeps latency low; a peering dispute is what makes it suddenly triple.
- Anycast
- The same IP address announced from many locations at once, so you reach whichever instance is closest. It is why a trace to a large DNS or CDN provider often ends surprisingly near you.
- MSS
- Maximum Segment Size: the largest chunk of data a TCP connection will send at once. When it is mismatched along a path, small requests work and large transfers stall — a failure that a plain ping will never reveal.