|This is an edited version of http://www.datastax.com/documentation/cassandra/2.0/cassandra/architecture/architectureIntro_c.html (Under Documentation > Home> Understanding the architecture > Architecture in brief ) for feedback purposes.
In this edited version, I tried to explain things by what they’re supposed to do, rather than how they work. Ben Slade
Architecture in brief
An overview of Cassandra’s structure.
Cassandra is designed to handle big data workloads across multiple nodes with no single point of failure. Its architecture is based on the understanding that system and hardware failures can and do occur. Cassandra addresses the problem of failures by employing a peer-to-peer distributed system where all nodes are the same and data is distributed among all nodes in the cluster [can a cluster be defined across data centers? yes]. Each node exchanges information across the cluster every second [same for multiple data centers?]. A [sequentially written] commit log on each node captures write activity to ensure data durability. Data is also written to an in-memory structure, called a memtable, and then written to a [append only] data file called an SSTable on disk once the memory structure is full [can the same data be in the memtable and the SSTable at the same time? Ie., can the memtable act like a cache?]. [Is the commit log truncated when data is written to the SSTable?] All writes are automatically partitioned and replicated throughout the cluster. Using a process called compaction Cassandra periodically consolidates SSTables, discards tombstones (an indicator that a column was deleted), and regenerates the index in the SSTable.
Cassandra is a row-oriented database. Cassandra’s architecture allows any authorized user to connect to any node in any data center and access data using the CQL language. For ease of use, CQL uses a similar syntax to SQL. From the CQL perspective the database consists of tables. Typically, a cluster has one keyspace per application. Developers can access CQL through cqlsh as well as via drivers for application languages.
Client read or write requests can go to any node in the cluster. When a client connects to a node with a request, that node serves as the coordinator for that particular client operation. The coordinator acts as a proxy between the client application and the nodes that own the data being requested. The coordinator determines which nodes in the ring should get the request based on how the cluster is configured. For more information, see Client requests.
Key components for configuring Cassandra¶
- Gossip: A peer-to-peer communication protocol to discover and share location and state information about the other nodes in a Cassandra cluster.Gossip information is also persisted locally by each node to use immediately when a node restarts. You may want to purge gossip history on node restart for various reasons, such as when the node’s IP addresses has changed.
- Partitioner: A partitioner determines how to distribute the data across the nodes in the cluster. Choosing a partitioner determines which node to place the first copy of data on.[Partitioners use various algorithms to assign the key value of a data row to an integer “token”.] You must set the partitioner type and assign the node a num_tokens value for each node (the more tokens assigned to a node, the more data will be stored there). [New configurations typically use virtual nodes to evenly spread tokens across (physical) nodes]. If not using virtual nodes (vnodes), use the initial_token setting instead.
- Replica placement strategy: Cassandra stores copies of data [a “copy of data” or “replica” is a group of data rows mapping to the same token value] on multiple nodes to ensure reliability and fault tolerance. A replication strategy determines which [on how many redundant] nodes to place replicas [within a data center, if applicable]. The first replica of data is simply the first copy; it is not unique in any sense.When you create a keyspace, you must define the replica placement strategy and the number of replicas you want. [The NetworkTopologyStrategy is highly recommended for most deployments because it is much easier to expand to multiple data centers when required by future expansion If using data centers, you define the number of replicas you want in each data center. Each data center holds a copy of all data].
- Snitch: A snitch defines the topology information [eg., racks and data centers] that the replication strategy uses to place replicas and route requests efficiently. [By default, the snitch software monitors the performance of reads from the various replicas and chooses the best replica for reading based on this history]You need to configure a snitch when you create a cluster. The snitch is responsible for knowing the location of nodes within your network topology and distributing replicas by grouping machines into data centers and racks.
- The cassandra.yaml file is the main configuration file for Cassandra. In this file, you set the initialization properties for a cluster, caching parameters for tables, properties for tuning and resource utilization, timeout settings, client connections, backups, and security.
- Cassandra stores table properties in the system keyspace. You set storage configuration attributes on a per-keyspace or per-table basis programmatically or using a client application, such as CQL.By default, a node is configured to store the data it manages in the /var/lib/cassandra directory. In a production cluster deployment, you change the commitlog-directory to a different disk drive from the data_file_directories.