1. There is no file for a given combination of parameters
    1. This can occur since the raw data does not provide sufficient granularity (early data only provided min/avg/max values for the 10 pings, later we provided the individual pings which enable calculation of IPDV for example) to provide certain metrics (e.g. jitter (IQR/IPDV), conditional packet loss, MOS), or a new metric was calculated (e.g. Directivity/Alpha, MOS) more recently  and we have not re-analyzed the older data to generate the new metrics.

  2. There is a file, but there is no data. I mean, file content only has the header (e.g., hours of a day) and dots (".")

    1. See item 3 below
  3. There is a file, there are some measures in the file, but for a specific  combination of parameters (say, in a specific hour of a day, for a given pair of nodes, and for a given metric) there was no data. That is, a dot is there.

    1. .'s are not unexpected. Usually they occur since the destination host is not reachable (e.g. down, changed its IP address, pings have been blocked to it, etc.). It can also be that the monitoring host was down (power outage, crashed etc.)  I pay attention when a remote host has not been seen for many months, then I check it's IP address has not changed (we actually ping the IP address rather than the name), whether the name is resolvable, whether the host responds to TCP but not ICMP echo etc.  and if necessary find another host to monitor that represents as similar country/city/community/affiliation. etc.  

  4. Handling Missing data in PingER and detecting Anomalies from Clusters by using Data Mining Techniques
  • No labels