buremba / hazelcast-modules

Hazelcast SPI modules

Hazelcast SPI Modules:

• HyperLogLog module
• TreeSet Module

You need to use these hazelcast configuration files or include the corresponding sections to your configuration in order to use the modules. Otherwise Hazelcast will not recognize service names.

• Server Config: src/resources/hazelcast.xml
• Client Config: src/resources/hazelcast-client.xml

HyperLogLog Module:

HyperLogLog module uses java-hll backend which is a Java implementation of HyperLogLog algorithm. Here is a sample usage:

HazelcastInstance server = Hazelcast.newHazelcastInstance();
hll = server.getDistributedObject(HyperLogLogService.SERVICE_NAME, "counter");
hll.add("TEST0");
hll.add("TEST1");
hll.add("TEST2");
long cardinality = hll.cardinality();

Currently supported features based on java-hll

• add
• addAll
• union
• cardinality

You can find more examples in test directory (src/test/java/org/rakam/cache/hazelcast/hyperloglog/HLLTest.java).

ITreeSet Module:

ITreeSet module uses native TreeSet data structure. You should implement Comparable interface to your classes, TreeSet keeps the the set in descending order. Therefore when you run treeset.iterate() it returns the ordered set. The insertion complexity of TreeSet is O(logN).

HazelcastInstance server = Hazelcast.newHazelcastInstance();
treeSet = server.getDistributedObject(TreeSetService.SERVICE_NAME, "set");
treeSet.add("TEST0");
treeSet.add("TEST2");
treeSet.add("TEST1");
for(Object item: treeSet) {
    System.out.println(item);
}

ITreeSet uses Hazelcast Collections so the supported methods are same as ISet.

You can find more examples in test directory. src/test/java/org/rakam/cache/hazelcast/treeset/TreeSetTest.java

####License

These Hazelcast modules is available under the MIT License.