Ignite Thread Pool usage - how do I properly put data in cache in parallel?

Question

As it is stated in docs, Ignite provides system thread pool to process cache related operations. My issue is a data putting, but streaming and batch loading is not a solution, because that data comes from multiple clients through REST-queries, and fast write acknowledge for each query is required.

So I'm looking for a way to make cache.put() operation in parallel. And first question is - is it true that simple cache.put() is not executed in parallel? I found out that I can invoke Ignite custom ExecutorService as following:

 ExecutorService exec = ignite.executorService();
 <...>
 exec.submit(new IgniteRunnable() {
      @Override public void run() {
        System.out.println("data put");
        cache.put(i, somedataobj);
                            }});

and then all the put operations are executed in parallel on all nodes (because I see "data put" in console output on all nodes). But I think it is not a proper way to make put in parallel, because data firstly sent to worker node, only then is put.

Another way is using Java plain fixed thread pool like that:

        ExecutorService exec = Executors.newFixedThreadPool(Runtime.getRuntime().availableProcessors());
        long startTime = System.currentTimeMillis();
        for (int i = 0; i < 10000; i++) {
            int t = i;
            DataClass dataobj = data().build();
            exec.submit(new IgniteRunnable() {
                            @Override public void run() {
                                System.out.println("data put");
                                cache.put(t, dataobj);
                            }});

        }

        exec.shutdown();
        exec.awaitTermination(0, TimeUnit.MILLISECONDS);
        long stopTime = System.currentTimeMillis();
        long elapsedTime = stopTime - startTime;
        System.out.println(elapsedTime);

But then I get another problem - data get loaded very fast (I get small elapsedTime), but appears in cache much later (when I do cache -a through ignite visor, I see how cache size is growing)

How do I properly put data in parallel? How can I use Ignite System Thread Pool to put data in parallel?

Answer 1

Just use async API, for example, IgniteCache.putAsync method: https://ignite.apache.org/releases/latest/javadoc/org/apache/ignite/IgniteCache.html#putAsync(K,%20V)