[GitHub] markusthoemmes opened a new pull request #2797: Use non-blocking computation of cache values.

[gi...@git.apache.org]

markusthoemmes opened a new pull request #2797: Use non-blocking computation of cache values.
URL: https://github.com/apache/incubator-openwhisk/pull/2797
 
 
   Currently, each time we try to get a value out of our cache `putIfAbsent` is called, which causes thread-contention. Instead, this uses `computeIfAbsent` which does a `get` first and thus is non-blocking for getting cached values under load.
   
   ## Testcode:
   
   ```scala
   object Test {
     import scala.concurrent.ExecutionContext.Implicits.global
   
     implicit val as = ActorSystem()
   
     def main(args: Array[String]): Unit = {
       val cache: ConcurrentMapBackedCache[String] = new ConcurrentMapBackedCache(
         Caffeine
           .newBuilder()
           .asInstanceOf[Caffeine[Any, Future[String]]]
           .expireAfterWrite(5, TimeUnit.MINUTES)
           .softValues()
           .build()
           .asMap())
   
       (1 to 100000000).par.foreach { _ =>
         cache.apply("testKey", () => akka.pattern.after(3.seconds, as.scheduler)(Future.successful("test2")))
       }
   
       as.terminate()
     }
   }
   ```
   
   ## With our current code:
   
   ![image](https://user-images.githubusercontent.com/1289147/30746992-12c4fd96-9fac-11e7-9f6d-1c238307df68.png)
   
   ![image](https://user-images.githubusercontent.com/1289147/30746995-16d39d7a-9fac-11e7-9d97-8e8fef738ac1.png)
   
   ## With the updated code:
   
   ![image](https://user-images.githubusercontent.com/1289147/30747000-1b6d21ee-9fac-11e7-8f77-3c9f10c83ad5.png)
   
   
 
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services