performance - Writing large amount of data using Akka in more efficient way -

i've implemented scala akka application streams 4 different types of data biomodule sensor (ecg, eeg, breath , general data). these data (timestamp , value) typically stored in 4 different csv files. however, have store each sample in 2 different files different timestamps, application writing in 8 different csv files @ same time. i've implemented 1 akka actor responsible persisting data, receive path file in write data, timestamp , value. however, bottleneck, since number of samples need store large (e.g. 1 ecg sample received each 4ms). result, actor had finished recording in short experiment 1-2 minutes after experiment over.

i've tried 4 actors 4 different message types, idea distribute work. didn't notice significant improvement in performances.

i'm wondering if has idea how improve performance. better use 1 actor storing files, few actors or efficient if have 1 actor each file? or maybe, doesn't make difference? improve code storing data?

this method responsible storing data:

  def processvalue(sample: waveformvalue): unit ={       val csvfilewriter=new printwriter(new bufferedwriter(new filewriter(sample.filepath,true)))       csvfilewriter.append(sample.timestamp.tostring)       csvfilewriter.append(",")       csvfilewriter.append(sample.value.tostring)       csvfilewriter.append("\r\n")       csvfilewriter.flush()       csvfilewriter.close() 

}

it seems me bottleneck i/o -- disk access. looks opening, writing to, , closing file each sample, expensive. suggest:

1. open each file once, , close @ end of processing. might need store file in member variable, or if have have arbitrary collection of files store them in map in member variable.
2. don't flush after every sample write.
3. use buffered writes each file writer. avoids flushing data filesystem every write, involves system call , waiting data written disk. see you're doing this, benefit lost since flushing/closing file after each sample anyway.