Database Reference
In-Depth Information
java.util.concurrent.ExecutorService
. To see specifically how this works, check out the
org.apache.cassandra.concurrent.StageManager
class.
The following operations are represented as stages in Cassandra:
▪ Read
▪ Mutation
▪ Gossip
▪ Response
▪ Anti-Entropy
▪ Load Balance
▪ Migration
▪ Streaming
A few additional operations are also implemented as stages. There are stages for units of work
performed on memtables (in the
ColumnFamilyStore
class), and the Consistency Manager is a
stage in the
StorageService
.
The stages implement the
IVerbHandler
interface to support the functionality for a given verb.
Because the idea of mutation is represented as a stage, it can play a role in both insert and delete
operations.
SEDA is a powerful architecture. Because it is event-driven, as its name states, work can be per-
formed with exceptional concurrency and minimal coupling.
Managers and Services
There is a set of classes that form Cassandra's basic internal control mechanisms. I'll present
a brief overview of these here so that you can become familiar with some of the more
important ones. The first one to consider is probably the
org.apache.cassandra.thrift.CassandraServer
class. This class implements the calls
to the Thrift interface, and delegates most of the efforts of performing queries to
org.apache.cassandra.service.StorageProxy
.
Cassandra Daemon
The
org.apache.cassandra.service.CassandraDaemon
interface represents the life cycle
of the Cassandra service running on a single node. It includes the typical life cycle operations
that you might expect:
start
,
stop
,
activate
,
deactivate
, and
destroy
.
