ADQM architecture
The high-level architecture view of ADQM is shown below.
Main points of the scheme:
-
To access ADQM, users can use JDBС and ODBC drivers, HTTP interface, console client, as well as wrappers on Python, PHP, Node.js, Perl, Ruby, Go, etc. Additional information about interfaces for working with ClickHouse is available in the Interfaces section of the ClickHouse documentation.
-
ADQM cluster includes the following services:
-
ADQMDB — ClickHouse column-oriented database management system. Data can be stored across different shards, where each shard is a set of replicas. Replication works at the table level on each shard independently of other shards.
-
ZooKeeper — is responsible for coordinating the process of data replication and is not involved in processing and executing queries. Alternatively, data replication can be organized by the ClickHouse Keeper service embedded in ADQMDB/ClickHouse.
-
Chproxy — an HTTP proxy and load balancer for the ADQMDB/ClickHouse database.
Users can manage an ADQM cluster using Arenadata Cluster Manager (ADCM) — a universal hybrid landscape orchestrator that allows installing, configuring, and upgrading Arenadata data services via the graphical interface.
-
-
ADQM can connect to an external data source (directly or via JDBC Bridge) and access its data. For example, ADQM can be integrated with the following systems: MySQL, MongoDB, Hadoop (HDFS), Apache Kafka, PostgreSQL, etc. The Table Engines for Integrations article lists all supported integrations.
For connecting to Arenadata products — Arenadata Database (ADB), Arenadata Hadoop (ADH), Arenadata Streaming (ADS), Arenadata Postgres (ADPG) — native integration is available within Arenadata Enterprise Data Platform (EDP). -
For visualizing data processed by ADQM and generating analytical reports, any BI tools providing access via JDBC/ODBC can be used. For example, BusinessObjects, Power BI, Luxms, Visiology, etc.