HeatSeeker

Understanding Data Temperature (Temperature being the frequency of access of the data) allows critical, informed decisions to be taken about the management of that data and the warehouse in which the data resides, allowing significant potential improvements in the TCO and RoI of the Data Warehouse itself. Consider what you could do if you that significant percentages of the data in the warehouse weren’t being used, or users were accessing key data only very infrequently, what percentage of data could be purged or archived if you had confidence in the data’s usage profile?

HeatSeeker is a data visualisation-based application, targeted specifically at analysing and managing Multi-Temperature data (Multi-Temperature data being data accessed with different frequencies) on Teradata data warehouses. As with other Ward Analytics applications, HeatSeeker has been designed and built from a clean sheet of paper, and directly address the issues encountered when trying to understand the usage characteristics of the data with a data warehouse;

HeatSeeker allows the user to simultaneously analyse thousands of databases, tables, columns, rows and the users accessing them, for a wide range of parameters and metrics that all can have an effect in isolation or combination on the “Temperature” of the warehouse data.

Data/Table/DB Perspective

A wide range of Metrics and Parameters are assessed in the analysis of the data temperature include (but are not limited to); table size, statistics, use of Multi-Value Compression, Indexes, Queries etc. all from a Data-User. System & Time Perspective.

Whilst the data itself is the initial point of focus when assessing Data Temperature, of equal importance are the users and the workload they generate in day to day usage of the warehouse. Some of the parameters key to understanding the behaviour of the warehouse from a User perspective are (but again not limited to); PSF and WLD groupings, Spool limits, # of Queries run by the Users, which tables the queries are being run against, Comment strings, # of Queries by PSF Group, # of Queries as FTS’s, # of Queries using single tables, Parallel Efficiency of the Queries etc.