Maria DB VS Raptor DB - Record Update Performance

[Forrest Falk]

My name is Forrest Falk, and I am a ServiceNow Certified Master Architect (CMA). I have had the opportunity to test both RaptorDB and RaptorDB Pro instances. With these tools, I have decided to start a series in the community where I will be testing different features of RaptorDB and RaptorDB Pro compared to MariaDB. This series should provide us with valuable insights into the advantages and disadvantages of RaptorDB and RaptorDB Pro in comparison to MariaDB. I will do my best to share the testing methods and results with everyone. If you have any questions, I would be happy to explain my testing process, and I encourage you to share any suggestions on what I should test next. I would love to keep the series interactive.

If you are just joining me, this is the second post in our series. Checkout the other posts below:

1. https://www.servicenow.com/community/developer-blog/maria-db-vs-raptor-db-record-insertion-performan...
   
   

For the second test, I want to focus on the raw speed of updating records using each database. In this test, we will measure how quickly each database can update records into the system. We will be updating the short description on the one thousand incidents we created in the first test as quickly as possible to see how long the process takes. Since we know the exact number of records, we can also calculate the average speed of each database across all one thousand records. I also plan to run multiple tests on the instances to account for any unrelated system activities such as scheduled jobs, scripts, or workflows that might affect the results.

Instance Version: glide-zurich-07-01-2025__patch0-07-15-2025

Below is the script I created to help test the update of incidents. The script measures how long it takes to update the short description on one thousand incidents. To make it easier to track these records, I included the phrase "Incident speed test V2" in both the short description and the description fields. I mainly did this to facilitate testing of record update and deletion performance with the same records.

Script to Update Records:

// Initialize timer
  var start = new GlideDateTime();
 
  // Query test incidents
  var gr = new GlideRecord('incident');
  gr.addEncodedQuery('short_descriptionSTARTSWITHIncident speed test V2');
  gr.query();
 
  // Updaterecords one by one
  var count = 0;
  while (gr.next()) {
    gr.short_description = gr.short_description + ' updated!';
    gr.update();
    count++;
  }
 
  // Capture end time and compute durations
  var end       = new GlideDateTime();
  var totalMs   = GlideDateTime.subtract(start, end).getNumericValue();
  var averageMs = totalMs / count;
 
  // Log results
  gs.print('FF: Total time to update' + count + ' incidents: ' + totalMs + ' ms');
  gs.print('FF: Average time per incident update: ' + averageMs.toFixed(2) + ' ms');
  gs.log('FF: Total time to update' + count + ' incidents: ' + totalMs + ' ms');
  gs.log('FF: Average time per incident update: ' + averageMs.toFixed(2) + ' ms');

Test Results:

Test 1

Maria DB:

*** Script: FF: Total time to update1000 incidents: 136314 ms

*** Script: FF: Average time per incident update: 136.31 ms

Raptor DB:

*** Script: FF: Total time to update1000 incidents: 50140 ms

*** Script: FF: Average time per incident update: 50.14 ms

Raptor DB Pro:

*** Script: FF: Total time to update1000 incidents: 50140 ms

*** Script: FF: Average time per incident update: 50.14 ms

Test 2

Maria DB:

*** Script: FF: Total time to update1000 incidents: 137615 ms

*** Script: FF: Average time per incident update: 137.62 ms

Raptor DB:

*** Script: FF: Total time to update1000 incidents: 48753 ms

*** Script: FF: Average time per incident update: 48.75 ms

Raptor DB Pro:

*** Script: FF: Total time to update1000 incidents: 48823 ms

*** Script: FF: Average time per incident update: 48.82 ms

Test 3

Maria DB:

*** Script: FF: Total time to update1000 incidents: 135010 ms

*** Script: FF: Average time per incident update: 135.01 ms

Raptor DB:

*** Script: FF: Total time to update1000 incidents: 48231 ms

*** Script: FF: Average time per incident update: 48.23 ms

Raptor DB Pro:

*** Script: FF: Total time to update1000 incidents: 48248 ms

*** Script: FF: Average time per incident update: 48.25 ms

To calculate how much faster RaptorDB and RaptorDB Pro were I used the following formula:

(Old Time / New Time)  = times faster

For each test, I used MariaDB as the baseline and then calculated how much faster RaptorDB and RaptorDB Pro performed in comparison. The fastest result was achieved in Test 3, where RaptorDB completed the task in 48231 milliseconds. The slowest result occurred with MariaDB, which took 137,615  milliseconds during Test 2. By reviewing the table below and averaging the results, we can determine the mean times for each database. Using these averages, we can also calculate how many times faster RaptorDB and RaptorDB Pro were throughout our tests.

Test #	MariaDB (ms)	RaptorDB (ms)	RaptorDB Pro (ms)
1	136314	50140	50140
2	137615	48753	48823
3	135010	48231	48248
Average Times:	136313	49041	49070
Times Faster than MariaDB:	1	2.78	2.78

After examining the results, we can observe that in this record update test, RaptorDB and RaptorDB Pro are approximately 2.78 times faster than MariaDB. However, it's important to note that this single test does not provide a comprehensive comparison between MariaDB and RaptorDB; it simply offers some insight into their raw performance. There was also an oddity in Test 1 where RaptorDB and RaptorDB Pro both had the same time of 50,140ms. Very rare to have the exact same time, but based on the other sites it seems their performance was similar for updating records.

The views, thoughts, and opinions expressed here belong solely to me, and not necessarily to my employer, organization, committee, or any other group or individual.