Ping ER Management Motivation n The Ping ER

Ping. ER Management

Motivation n The Ping. ER database stores ping results from a huge number of monitoring site -remote site pairs. n The interface of Ping. ER just presents all this data, with no distinction between down sites , sites showing anomalous rtt values etc. n An administrator has to see everyday – whether all sites and nodes are up ? – Are there any anomalous sites and nodes ? – Is the data collected by monitoring sites valid or not ?

Proposed Solution n This system would automatically – Store the latest results regarding down sites and down nodes on the web – Find out Anomalous sites – Find out invalid data and mismatch TLD’s (Top Level Domains) in the Ping. ER database – Make various histograms and pie charts to help the Administrator to find the current status of the Ping. ER system – Optionally Email the current status of the sites and nodes to the Administrator

Problems with previous implementation n Storage of too much redundant information. n Very inefficient database design – Too many databases and tables – (6 databases and 30 tables) – Same information stored in multiple tables

New design n Total of 5 tables , each concerned with a separate and independent issue : – – – Downsites Dramatic rtt Anomaly Invalid Data NOT-Set Invalid TLD

Done n Downsites Time Stamp Mon_ Site Rem_ Site Status Down days Mon_ Country Rem_ Mon_ Country Region Rem_ Region

Done n Mon_ Site Dramatic rtt Anomaly Rem_ Site Minrtt Sdrtt Time Stamp Mon_ Country Rem_ Country Mon_ Region Rem_ Region

To be done NOT- Set n Invalid TLD n Invalid Data n

Documents n http: //pinger 2. niit. edu. pk/apache 2 -default/Pingermanagement/index. html