It's all about the answers!

Ask a question

Will the snapshot impacts the performance of RTC

Hehui Liu (7052830) | asked Jan 14 '13, 9:58 a.m.
Here, we want to automatically create one snapshot each time a change set is delivered,
but we do not know whether this will impact the performance of RTC.
Could any one give advise on this? Thanks a lot!

One answer

permanent link
Andrew Hoo (1.0k1) | answered Jan 14 '13, 11:25 a.m.

While this article is talking about Repository Workspaces in particular, Snapshots should be using the same backing data as well.

So the answer is "No, having lots of snapshots should not impact performance because RTC was designed with scalability in mind"

That being said though, I think you should evaluate why you want to do this. Like, what are you doing with snapshots (each with only 1 change set of incremental difference than the last) that cannot be done by examining the history of the components in a stream? (Since you seem to already be assuming that change sets are being added one at a time).

Even if only a small amount of data is used for each new Snapshot, if you do it A LOT, then the small amounts will add up.

Also, you might run into user workflow issues. If a dialog pops up asking you to pick a snapshot and you see hundreds of snapshot per day, then as a human day-to-day user, how do you decide what you really want?

Geoffrey Clemm commented Jan 20 '13, 9:20 p.m.

I've had customers ask about something like this because of two reasons:
- The history is being maintained by RTC at the component level, but not at the full stream level.   A snapshot is needed to maintain history at the full stream level.
- The change-set level history of a stream can be re-ordered by a deliver.  A baseline is needed at each deliver to maintain the history of the stream at the change-set level.

Your answer

Register or to post your answer.