how the value of 'thresholdPermits' is cacluated in the SmoothWarmingUp?

[张凯]

https://github.com/google/guava/blob/master/guava/src/com/google/common/util/concurrent/SmoothRateLimiter.java#L226

shows the formula of the 'thresholdPermits'

i cannot understand why doing so, why not 0.2 * warmupPeriod/ stalbeInterval

thanks very much!

[Dimitris Andreou]

It's explained few lines above, L193 in particular. Basically, the code implements a mathematical function which consists of two lines. The two lines meet at a point ("thresholdPermits"). On the left of that point, the function is horizontal line, on the right, it's a sloped line (the sloped line corresponds to the warming up effect). The code you linked recomputes that meeting point, given that it depends on the rate.

--
guava-...@googlegroups.com
Project site: https://github.com/google/guava
This group: http://groups.google.com/group/guava-discuss
 
This list is for general discussion.
To report an issue: https://github.com/google/guava/issues/new
To get help: http://stackoverflow.com/questions/ask?tags=guava
---
You received this message because you are subscribed to the Google Groups "guava-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to guava-discus...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/guava-discuss/65abb755-9efb-4c60-bc33-e7cf703fccce%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[张凯]

OK, thank you

i am puzzle with the sentence on line https://github.com/google/guava/blob/master/guava/src/com/google/common/util/concurrent/SmoothRateLimiter.java#L181

"The reason that this is warmupPeriod/2 is to maintain the behavior of the original implementation where coldFactor was hard coded as 3"

so the value of warmupPeriod/2 is empirical valued ？

and can i choose the time to go from thresholdPermits to 0 with another vaule, such as warmupPeriod/3

在 2018年10月11日星期四 UTC+8下午10:05:46，Dimitris Andreou写道：

[Dimitris Andreou]

It's an arbitrary value rather, not empirical. Most people who want a warmup behavior (and that's already very few people) don't care about the precise shape of the throughput during a warm-up or cool-down, so it doesn't make sense to offer an overcomplicated API that nobody would use.

This particular value controls the cool-down effect of a previously fully-utilized ratelimiter which is now unused. If this value was shorter (even 0), then the ratelimiter would immediately start to cool-down as soon as it is not fully utilized. That the value is >0 means that there's a buffer, a fully utilized ratelimiter doesn't start cooling down (thus throttling more, to warmup) just because there was a blip in incoming traffic.

To view this discussion on the web visit https://groups.google.com/d/msgid/guava-discuss/549571e1-3509-487a-9fcb-6744e9e914e2%40googlegroups.com.

[张凯]

i got it, thank you ;)

在 2018年10月12日星期五 UTC+8下午5:05:39，Dimitris Andreou写道：