Today I'm gonna share my experience when I started with Apache Cassandra... If you consider the case study of Digg: How Digg migrated their friends+diggs data set to Cassandra from MYSQL, here is the explanation.
Digg implemented their new badges feature in Cassandra, a distributed, structured key-value store. Digg adopted a traditional vertically distributed master-slave configuration with MySQL, and investigated sharding (type of database partitionin) MySQL. In the case of the traditional architecture, Cassandra has a masterless "ring" which is easy to set up and to maintain.
Since it was necessary to strand data normalization and consistency to make these methodologies work. Digg adopted to comfortable with non-relational data stores. With the consideration of HBase, Voldemort, Cassandra, Hypertable, Tokyo Cabinet/Tyrant, and Dynomite, finally migrated on to Cassandra.
Every system has its own pros and cons, but Cassandra has a good blend of all. It offers Data Partitioning, Consistent Hashing, Data Replication, Eventual & Tunable Consistency, Gossip Protocol to discover nodes, Keyspace, Column & Row -oriented data storage, so you have a bit more structure than plain key/value stores. It operates in a distributed, highly available, Cluster/Ring.
When an item is dugg, Digg asynchronously populates Cassandra. This operation fetches the list of followers of the digging user, and every column is inserted at once, atomically in each of their buckets, under a second.
In my upcoming posts, I will try to explain Cassandra architecture and other topics.جميع الحقوق محفوظة