- Core Distributed in memory storage using consistant hashing
- Node Add/Remove in runtime with data realocation
- Data Replication with Dynamic/Adaptive Replication
- Failure detection
- Cluster Discovery
- Consensus protocol
- Performance analysis
- Make Raft reusable with objects
import anjanas/distributed_cache as cache;
cache:connectToCluster();
cache:Cache oauthCache = new("oauthCache");
oauthCache.put("Name", "Ballerina");
io:println (<string>oauthCache.get("Name"));If you just want to see distributed cache in action in cluster mode you can simply execute run the following shell
script to spawn processes without doing any configurations yourself
./ThreeNode.sh
[cache]
ip="http://localhost" #IP of the Node
port=7000 #Port for cache communication
hosts="" #IP and ports of the known nodes in the cluster
capacity=10 #Entry capacity for the node
request.timeout=2000 #Timeout of remote calls
relocation.timeout=60000 #Timeout for a relocation request
eviction.factor=0.1 #Amount of entries will be cleaned in a cleanup
replication.fact=1 #Replicas available of one entry
[local.cache]
enabled=false #Local cache availability
capacity=100 #Capacity of the local cache
eviction.factor=0.25 #Eviction factor of the local cache
[raft] #Keep these at default values if you don't know about Raft
min.election.timeout=2000 #Maximum election timeout for raft
max.election.timeout=3000 #Minimum election timeout for raft
heartbeat.timeout=1000 #Heartbeat timeout for raft
[failure.detector]
suspect.value=10 # Suspect points increased for each connection failure
timeout.millis=1000 # Timeout for failure detector
backoff.fact=1 # Backoff factor for failure detector
[consistent.hashing]
partitions=7 # Number of partitions in the hash ring