The current request is to have buckets of cache so when multiple people are looking at specific data statically we have a manager to tell the code that something within a bucket changed and to reload other people's screen. The idea here is to look at a smaller table so that we are not querying the larger tables to see when things change. This is extra overhead.
On a Save sproc now we have to update the sync manager when we save the record. Seems a bit wasteful to me. We will loose performance on the save and potentially get locks on that table. I don't see another avenue though. Has anyone done a project that I can pull off of on this?
Edit: This will not be a website. This is a windows form app that will probably support 20-40 concurrent users. I would think 40 would be max. The tables are heavy. Meaning that we could be looking at gigs of data. Better if I speak to the number of rows. The tables they want to know if the cache is in sync will have roughly a 4 million rows within a year of application start up. Here is a website I found that describes exactly what I am after
They go into the benefits and cons a bit, but no real context for data models.
The model that we are coming up with is similar to this
A Sync Manager table
PK int sync Manager Id
Snyc Manager type smallint which is the entity within the bucket. So each entity will have there own ID
Owner ID uniqueidentifier (required datatype to match existing data structure) This is what it was opened from. Where in the tree it was opened to know the control level
TimeStamp - This will need to be updated whenever there is a change to know that something in cache is out of sync and needs to be refreshed
It is really managing two tables all the time. the OLTP table and the Sync table to make sure the cache is up to date
Hopefully this helps a bit