Sumerian’s data collector captures configuration and usage data from your existing infrastructure monitoring and application performance management tools. This is typically configured to run automatically, every night.
The captured data is compressed and encrypted, and pushed to our data centre for processing.
This way you benefit from our big data analytics with a minimal deployment footprint on your premises.
For most capacity management operations you want to understand how resource consumption varies statistically, so the time series data for all resources is analysed, and a statistical profile created.
The raw time series data is retained for those times when you absolutely need it.
The most immediate question for capacity management is to determine the headroom for different types of demand in various parts of the estate.
This is automatically calculated by comparing the statistical demand for resources, with the resources available.
Resource consumption metrics are used to train a series of machine learning algorithms. If the trained models pass a series of automated quality tests they are used to generate future time series estimates. Any future consumption which exceeds a configurable threshold is then exposed as a Capacity Event.
Correlations between application volumes, application metrics and resource utilisations are continuously mined from multiple data sources, then encoded and made available to use within the product, allowing for the projection of utilisation and performance based on planned business growth.
Baselines and Scenarios
Your baselines are now updated and available for reporting and modelling.
Any scenarios connected to the updated baselines are also updated, so you continually get the most accurate view of your future.