Pigeon Voice Server Database Log Disk Usage - DatasmithSA/Pigeon-Voice-Training GitHub Wiki

When having the Log To Database setting enabled, the Pigeon Voice Server will generate a substantial amount of data in its PigeonVoice SQL Server database. For this reason it is important to take the database disk usage into consideration when deciding on disk space for the server that will be hosting the Pigeon Voice Server.

There are many factors that determine the database disk usage/growth:

  • Number Of Operators the more operators are working in the warehouse the more database log entries you can expect.
  • Number of Hours in a Shift: the number of hours operators work for in a single shift.
  • Number of Shifts / Day: some warehouses have both day and night shifts, while others only operate on in a single shift.
  • Number Of Messages / Hour : the more messages are sent/received by the Pigeon Voice Server the quicker and larger the database log will grow. This could be based on:
    • How much work is assigned to each operator i.e. how many picks they have to complete in a day/hour etc.
    • Complexity of the voice dialog: dialogs that have long prompts and many prompts and answers will obviously generate less messages with the server due to pickers talking more instead of working.
    • Voice Architecture: some voice implementations are designed to work completely live while others cater for off-line functionality. Some Voice applications download a complete assignment (e.g. pick list/order) to the device in a single transaction and thereafter send the pick confirmations back to the server in separate transactions. These type of applications will require less download transactions between the server and devices as opposed to applications designed to download and upload each pick confirmation one at a time i.e. batching of data can reduce the amount of transactions and therefore the database disk usage.
      • The number of message types (calls to Pigeon Voice Agents) and the frequency of the messages between the devices and the server can also have an impact on the message count and therefore database disk usage e.g. if there are additional messages required to notify the WMS of cartons being being opened/closed etc.

It is for all the above reasons that estimating the disk usage is not a straight forward task. It is therefore recommended that several trial runs be performed with each new voice implementation before go-live.

Example: the following serves as an example of the typical exercise you should follow when determining the hard disk requirements for your server that will be hosting the Pigeon Voice Server. In this example, several trial runs were gone through in a live environment with several voice operators and devices connecting to the Pigeon Voice Server concurrently.

You can use the results of these trial runs as a guideline.

We began with an empty SQL Server 2012 Express database which is limited to a size of 10GB.

  • Test 1: 75 minutes 20 pickers
  • Test 2: 120 minutes 30 pickers
  • Test 3: 40 minutes 17 pickers

Average pickers: (20 + 30 + 17) / 3 = 22 pickers on average during test. Total minutes: 235

Calculating Current Disk Usage:

  • Database Disk Usage Total (in 235 minutes): 73MB
  • Database Disk Usage / Minute: 73MB / 235minutes = 0.3106MB per minute (at 22 pickers/devices)
  • Database Disk Usage / Minute / Per Picker: 0.3106MB / 22 pickers = 0.0141MB (per picker / minute)

Calculating Expected Database Disk Usage in production:

  • Total Expected Pickers/Devices in your in production: 36
  • Database Disk Usage / Minute (at 36 pickers) = 0.5083MB
  • Database Disk Usage / Hour (at 36 pickers) = 0.5083MB * 60 = 30.5MB
  • Database Disk Usage / Day (8 hour shift at 36 pickers) = 30.5MB x 8 = 244MB
  • Maximun size of SQL Server 2012 Express Database: 10GB = 10000MB
  • Days to reach Database Disk Usage Limit (at 36 pickers): 10000MB / 244MB = 41 days
  • Hours to reach Database Disk Usage Limit (at 36 pickers): 10000MB / 30.5MB = 327 hours

In order to manage the size of the Pigeon Database Log, the next section will cover Database Log Purging. This is a feature provided by the Pigeon Voice Server in order automatically purge (delete) historical log entries.