Page 70 - Building Big Data Applications

64 Building Big Data Applications

user interaction and keep the consistency, Cassandra provides a model called tunable
consistency. In this model, the client application decides the level of consistency desired
by that application. This allows the user the flexibility to manage different classes of
applications at different levels of consistency. There are additional built-in repair
mechanisms for consistency management and tuning. A key point to remember is
consistency depends on replication factor implementation in Cassandra.

Write consistency
Since consistency is a configuration setting in Cassandra, a write operation can specify
its desired level of consistency. Cassandra lets you choose between weak and strong
consistency levels. The following consistency levels are available.
Consistency
level Write consistency Read consistency
ANY A write must be written to at least one node Not applicable
If a replica is down, a live replica or current node
can store a hint and update the node when it
comes back live
If all replica nodes for the given rowkey are
down, the write can still succeed once has been
written by storing the hint and the data in the
coordinator
Not a preferred model
Notedif all replica nodes are down an ANY
write will not be readable until the replica
nodes for that rowkey have restored
ONE A write must be written to the commit log and Returns a response from the closest replica
memory table of at least one replica node
QUORUM A quorum is defined as the minimum number of replicas that need to be available for a successful
read or write
Quorum is calculated as (rounded down to a whole number): (replication_factor/2) þ 1
For example, with a replication factor of 5, a quorum is 3 (can tolerate 2 replicas down)
A quorum is a middle ground between weak and strong consistency
A write must be written to the commit log and memory table on a quorum of replica nodes to be
successful
Local quorumda write must be written to the commit log and memory table on a quorum of replica
nodes in the same data center as the coordinator node
Each quorumda write must be written to the commit log and memory table on a quorum of replica
nodes in all data centers
ALL A write must be written to the commit Data is returned once all replicas have responded.
log and memory table on all replica The read operation will fail if a replica does not
nodes in the cluster for that row key respond.
The highest and strongest level of
consistency brings latencies in the architecture
Not a preferred method until complexity and
volumes are low