Situation

We've got one server which stores data for multiple remote field devices. Each device has its own version of the database, and will try to sync whenever it comes into 3G coverage.

Each device only needs to know its own data, however it will need to download changes if the data is modified on the server after it has been uploaded.

Sending the whole database is not viable due to limited data connectivity, so we only what to send the rows that have changed.

Current Solution

Device stores datetime of when each table was last modified, compares with server, uploads and downloads rows modified or inserted after that datetime.

Problem

This solution seems nasty. We're wondering if there is a standard pattern for solving this problem that we can't seem to find information on?

有帮助吗?

解决方案

You can use merge replication with parametrized row filters. The basic idea is to create a publication of some articles (= tables) and then create a list of subcriptions. Here it is a T-SQL script that allows you to understand the concept.

许可以下: CC-BY-SA归因
不隶属于 StackOverflow
scroll top