Seamless data replication has emerged as a critical necessity for businesses striving to stay agile and competitive. Yet, the challenge remains: how can organizations harness the power of data replication without overhauling their existing architectural frameworks? Many organizations aren’t interested in or in a position to replace their centralized data stores.

Here, we focus on leveraging Azure Functions and HarperDB API to replicate data to the edge, which offers the following benefits:

• Data origin and existing systems unimpacted
• Offload system of record
• Reduced latency
• Continuous data sync

Implementation

The implementation is something like this:

• The Azure Function is executed through SqlTrigger when a change in a table is detected
• The code sends the corresponding updates to Harper DB

According to the documentation, this is the way the trigger works:

while (true) { Get a list of changes on the table – up to a maximum number controlled by the Sql_Trigger_MaxBatchSize setting. Trigger function with a list of changes Wait for delay controlled by the Sql_Trigger_PollingIntervalMs setting }

For example, let’s say we have a table called product:

and we want to synchronize the changes to our distributed DB:

For this to work, we must enable the change-tracking functionality in the database and the table.

The function uses the SqlTrigger:

[FunctionName(“HarperDBReplication”)] public static void Run( [SqlTrigger(“[SalesLT].[Product]”, “connectionStringSetting”)] IReadOnlyList<SqlChange> changes, ILogger log) {

On it, we specify the table we target and the corresponding connection string as a decorator for a List of SqlChanges of an entity called Product. This entity describes the fields we’d be retrieving

public class Product { public string ProductID { get; set; } public string Name { get; set; } public string Color { get; set; } public double ListPrice { get; set; } }

Whenever we make a change to the central database, the Azure Function is executed, sending the changes to Harper:

And the changes are reflected afterward:

Conclusion

This lightweight solution provides an efficient application replication mechanism, helping answer “What rows have changed for a table?” and ensuring data is up to date at the edge. The change tracking feature is supported in SQL Server 2016 (13.x) and above, Azure SQL and Azure SQL Managed Instance.