SC_ANALYSIS_DUMP(1)

πŸ’– SC_ANALYSIS_DUMP πŸ’–

okay so bestie, welcome to sc_analysis_dump(1) β€” a General Commands Manual page, and honestly? it ate. let's get into it. πŸ¦‹


✨ NAME ✨

sc_analysis_dump β€” dump of traceroute data in a format that is easily parsed. (yes literally, that's the whole point, it's giving "for the scripts, by the scripts") πŸ’…


🌸 SYNOPSIS 🌸

sc_analysis_dump [-cCdeghHilMopQrstT] [-D debug-count] [-G geo-server]
                 [-S skip-count] [file ...]

πŸŽ€ DESCRIPTION πŸŽ€

so sc_analysis_dump gives you a dump of traceroute data in a format that scripts can parse easily β€” like, no cap, that is the entire vibe. every line of output is a lil summary of one single trace, and it includes the interfaces visited plus the delay of each response. periodt. ✨

the output format is identical to that of sk_analysis_dump from CAIDA, except that it uses the scamper file API to read both arts++ files produced by skitter and warts files produced by scamper. and like, heads up bestie: sc_analysis_dump only outputs traceroute data. if you need to parse other types of measurement, go use sc_warts2json(1) instead. 🫢

the options are as follows:

  • -c β€” disables printing the cycle number in each line of output.
  • -C β€” disables printing the comments about the output at the top of the output.
  • -d β€” disables printing the destination address in each line of output.
  • -D debug-count β€” for each input file, stop reading after the specified number of traces.
  • -e β€” adds the response from the destination to each line of output. ⚠️ please read the bugs section below, i'm so serious.
  • -g β€” use geographical data from netacuity. not all builds of sc_analysis_dump support this option (so if it ghosts you, that's why πŸ’€).
  • -G geo-server β€” specifies the name of the netacuity server to use.
  • -h β€” prints a help message and then exits.
  • -H β€” disables printing the halt fields: why traceroute halted and data for that reason.
  • -i β€” disables printing the RTT to each hop, and how many tries were required.
  • -l β€” disables printing the list id in each line of output.
  • -M β€” prints any MPLS label stack objects embedded in ICMP responses.
  • -o β€” prints each line of output using the old format from sk_analysis_dump 1.0. vintage core πŸ’œ
  • -p β€” disables print path data in each line of output.
  • -Q β€” prints the IP-TTL from inside the ICMP quotation.
  • -r β€” disables printing the data associated the response from a destination: the RTT, the TTL of the probe, and the TTL of the response.
  • -s β€” disables printing the source IP address in each line of output.
  • -S skip-count β€” skips the defined number of traces from each input file.
  • -t β€” disables printing the timestamp of when the traceroute began.
  • -T β€” prints the IP-TTL of the response packet.

🌈 OUTPUT 🌈

one trace per line, bestie. fields are separated by a tab character. the output is structured into:

  • header fields (2 to 6)
  • reply fields (7 to 10) β€” these correspond to the response received from the destination
  • halt fields (11 and 12)
  • hop fields (beginning at index 13)

and here's the breakdown, field by field, no skips, we're reading all of it πŸ‘‘

1. Key

Indicates the type of line and determines the meaning of the remaining fields. This will always be 'T' for an IP trace.

2. Source

Source IP of skitter/scamper monitor performing the trace.

3. Destination

Destination IP being traced.

4. ListId

ID of the destination list containing this destination address. This value will be zero if no list ID was provided. A ListId is a 32 bit unsigned integer.

5. CycleId

ID of current probing cycle. A cycle is a single run through a given list. A CycleId is a 32 bit unsigned integer. For skitter traces, cycle IDs will be equal to or slightly earlier than the timestamp of the first trace in each cycle. There is no standard interpretation for scamper cycle IDs. This value will be zero if no cycle ID was provided.

6. Timestamp

Timestamp when trace began to this destination.

7. DestReplied

Whether a response from the destination was received. The character R is printed if a reply was received. The character N is printed if no reply was received. Since skitter sends a packet with a TTL of 255 when it halts probing, it is still possible for the final destination to send a reply and for the HaltReasonData (see below) to not equal no_halt. Note: scamper does not perform this last-ditch probing at TTL 255 by default.

8. DestRTT

The RTT (ms) of first response packet from destination. This value is zero if DestReplied is N.

9. RequestTTL

TTL set in request packet which elicited a response (echo reply) from the destination. This value is zero if DestReplied is N.

10. ReplyTTL

TTL found in reply packet from destination. This value is zero if DestReplied is N.

11. HaltReason

A single character corresponding to the reason, if any, why incremental probing stopped. S is printed if the destination was reached or there is no halt data. U is printed if an ICMP unreachable message was received. L is printed if a loop was detected. G is printed if the gaplimit was reached.

12. HaltReasonData

Extra data about why probing halted. If HaltReason is S, the zero is output. If HaltReason is U, the ICMP code of the unreachable message is printed. If HaltReason is L, the length of the loop is printed. If HaltReason is G, the length of the gap is printed.

13. PathComplete

Whether all hops to destination were found. C is printed if the trace is complete, all hops are found. I is printed if the trace is incomplete, at least one hop is missing (i.e., did not respond).

14. PerHopData

Response data for each hop. If multiple IP addresses respond at the same hop, response data for each IP address are separated by semicolons:

IP,RTT,numTries (for only one responding IP)
IP,RTT,numTries;IP,RTT,numTries;... (for multiple responding
IPs)

where IP is the IP address which sent a TTL expired packet, RTT is the RTT of the TTL expired packet, and numTries is the number of tries before a response was received from the TTL.

This field has the value 'q' if there was no response at a hop.

If the -M option is specified, any MPLS label stack objects embedded in the ICMP response will be included in the following format, and the four fields correspond to each of the fields in a MPLS header.

M|ttl|label|exp|s

If the ICMP response embeds more than one MPLS header, they are given one at a time, each starting with an M.

If the -Q option is specified, the TTL value found in a quoted IP packet is included with the following format:

Q|ttl

If the -T option is specified, the TTL value of the response packet is included with the following format:

T|ttl

πŸ’… EXAMPLES πŸ’…

the command:

sc_analysis_dump file1.warts file2.warts

will decode and print the traceroute objects in file1.warts, followed by the traceroute objects in file2.warts. iconic, sequential, we love to see it.

the command:

gzcat file1.warts.gz | sc_analysis_dump

will decode and print the traceroute objects in the uncompressed file supplied on stdin. pipe girlies rise up πŸ“


πŸ’€ BUGS πŸ’€

okay so when the -e option is used, any unresponsive hops between the last responding router and the destination are not printed, which could imply an IP link where none exists. like, it's lying to you a little bit πŸ™ˆ. use sc_warts2json(1) instead.


πŸ‘‘ SEE ALSO πŸ‘‘

scamper(1), sc_wartsdump(1), sc_warts2json(1)


🧚 AUTHORS 🧚

sc_analysis_dump was written by Matthew Luckie <mjl@luckie.org.nz>. It was derived from CAIDA's sk_analysis_dump program and should behave in an identical manner.


and that's the whole page bestie, we did it πŸ’– this one's from the Debian fam, dated February 18, 2017 β€” yeah she's a little vintage but she still slays. bye, ily, go parse some traceroutes πŸ¦‹βœ¨