The objective of the Traffic, Availability and Performance (TAP) index is to give application & program managers, business owners and solution stake holders, a less-technical orientated metric on their servers. FusionAnalytics makes suggestions on how to improve your TAP score, as well as measuring and comparing TAP performance over time.
TAP Overview
TAP provides a daily application score, consisting of the TAP Score (out of 10 – 10 is the highest) and TAP Grade (A..F – A is the best). The TAP Score and Grade are calculated on a daily basis and are based on various application metrics. TAP gives you a standardized comparison of applications regardless of function or hardware capability. The TAP score is built from 5 equally important metrics, broken down into 3 different categories:
- Traffic
- Number of impressions
- Number of sessions
- Availability
- Uptime percentage
- Restart count
- Performance
- Average execution time
Benefits
- TAP provides a single point of measurement to review how well your server is performing
- No technical knowledge required to understand TAP – score simplified to A, B, C, etc. grading
- TAP automatically scales to be relevant for your platform, no matter it’s size.
- TAP automatically re-adjusts over time to compensate for changes in environment and/or traffic patterns (e.g. more memory, new code deployment, configuration changes, “Slashdot effect” etc.), ensuring that TAP maintains relevance.
- TAP Report is emailed directly to you – no need to login to FusionAnalytics
- TAP Report includes achievable recommendations on how you can improve your score
- TAP Report compares historic data
- Each of last 7 days
- Weekly (Monday to Sunday)
- Same day in previous weeks (up to 5 previous weeks)
- TAP Report analyses trends for varying periods
- Week to date
- Last 30 days
- Year to date
- Day of week (eg last 5 Thursdays)
- TAP Reports are archived for later reference even after source data expunged from analytics server
How is TAP Calculated?
Each metric is equally important in the calculation which has a maximum of 10, thus each metric is worth up to 2 "TAP points".
- Traffic
- Number of impressions
- Calculated on a normalized scale from 365 days of historic data and then scaled between -4 and +4 standard deviations from the mean.
- +4 standard deviations = 100%
- -4 standard deviations = 0%
- Number of sessions
- Calculated on a normalized scale from 365 days of historic data and then scaled between -4 and +4 standard deviations from the mean.
- +4 standard deviations = 100%
- -4 standard deviations = 0%
- Availability
- Process uptime
- Calculated on a set exponential scale:
- No downtime = 100%
- 1 minute downtime = 97%
- 2 minutes downtime = 94%
- 3 minutes downtime = 91%
- 4 minutes downtime = 89%
- 5 minutes downtime = 86%
- 10 minutes downtime = 74%
- 20 minutes downtime = 55%
- 30 minutes downtime = 41%
- 60 minutes downtime = 17%
- Restart count
- Calculated on a set exponential scale:
- 0 restarts = 100%
- 1 restart = 90%
- 2 restarts = 81%
- 3 restarts = 74%
- 4 restarts = 67%
- 5 restarts = 60%
- 10 restarts = 37%
- 20 restarts = 14%
- 30 restarts = 5%
- Performance
- Average execution time
- Calculated on a normalized scale from 365 days of historic data and then scaled between -4 and +4 standard deviations from the mean.
- As lower execution times are better, the scaling used is the inverse of that for impressions and sessions.
- -4 standard deviations = 100%
- +4 standard deviations = 0%
