Redis in-memory read replica of an RDS master

Question

I have a high traffic web service. Each request SELECTs some data from an MSSQL RDS. The relevant tables in the RDS are very small, update/inserts will only happen few times a day.

In order to minimize each request time, I wanted to have some sort of an in-memory read replica Redis that will store those table in some sort of a key-value structure that I'll define in advance. And to make sure that every time an Update/Insert is done on the RDS, it will update the Redis. This application runs on multiple linux instances, so I'll need to have a way to deploy it to all of them, and obviously to keep them sync.

1.Does it make sense? maybe I should take a different approach?

2.If it does, can anyone share how would I achieve such thing? Is there an implemented tool in MSSQL/Redis or do I have to write some sort of a third-party app for it? couldn't find anything similar online.

Thanks for reading!

How do you sync data whenever an update/insert takes place to relevant table? Is that the issue you are finding a solution? — Chathuranga Ranasinghe, Feb 17 '20 at 12:33
any possibility for you to use single centralized Redis instance for all application servers? — Chathuranga Ranasinghe, Feb 17 '20 at 12:49
Nope... the app is heavily RDB.. just for this specific request which happens all the time I thought that storing the data in Redis could make it faster. — Ben, Feb 17 '20 at 12:51
You will get some idea if you go through this https://www.slideshare.net/RedisLabs/redisconf18-techniques-for-synchronizing-inmemory-caches-with-redis — Chathuranga Ranasinghe, Feb 17 '20 at 13:00

Redis in-memory read replica of an RDS master

0 Answers0