Arenadata Grid

Analytical platform with open source

The in-memory computing platform allows you to significantly speed up your applications without having to replace existing DBMS. Arenadata Grid (ADG) provides the ability to flexibly scale the resulting solution into dozens and hundreds of compute nodes. Using RAM as the primary drive, the Arenadata Grid platform is able to execute and process requests from 1,000 to 1,000,000 times faster than traditional disk DBMS. The computing power of the Arenadata Grid is easily scaled by adding new nodes to the cluster without the need to stop working, thereby allowing to process hundreds of terabytes of data from multiple databases.

You can easily upgrade the architecture of your existing applications by using ADG as an intermediate caching storage layer. The Arenadata Grid platform is able to transparently and seamlessly integrate with existing RDBMS, NoSQL DBMS and Hadoop. ADG implements a single API that supports SQL, MapReduce, and programming languages Java, C++, .NET, PHP. Arenadata Grid, along with the applications that use it, can be easily run both in the cloud and in your data center.

ADG can also be used as a distributed transactional DBMS with support for SQL queries, capable of storing data both in memory and on disk. This is achieved by using persistent storage technology in conjunction with the capabilities provided by the built-in SQL engine.

1,000x Better Performance Versus Disk-Based Databases

The in-memory computing platform Arenadata Grid allows you to increase the performance of used traditional DBMS by more than 1,000 times by reducing the cost of reading and writing from disk drives. Acting as an intermediate caching storage layer, ADG pushes data from your RDBMS or NoSQL DBMS to RAM. In addition to high-performance caching layer, Arenadata Grid implements a number of features, such as streaming data processing, distributed computing, Spark and Hadoop acceleration.

Arenadata Grid
Cluster monitoring is provided by Grafana/Graphite

The Arenadata Grid platform is able to scale almost linearly by adding new nodes to an existing cluster. The data in the cluster is optimally located to minimize data migration as the cluster topology changes. At the same time data migration is performed automatically.

One of the features of Arenadata Grid is the lack of a master node. The desired node that stores the target record is determined by calculating the value of the special function.

ADG implements a distributed key-value cache located between your applications and the DBMS, thereby allowing seamless integration into your existing architecture by performing end-to-end reading and writing.

Main advantages

  • Distributed cache in memory.
  • Lightning fast computing speed.
  • Flexible scalability.
  • Distributed transactions.
  • Distributed SQL queries.
  • Multi-level data storage (RAM/Disk) off-heap.

Arenadata Grid
ADG allows you to save data to third-party DBMS, providing end-to-end reading and writing

Store and Process Data In-Memory with ACID Transactions and ANSI-99 SQL Support

The Arenadata Grid platform supports ANSI-SQL 99. You can easily perform queries on an array of data distributed among hundreds of nodes that store data both in memory and on disk, paging them from third-party sources if necessary. You also have the opportunity to connect to Arenadata Grid from your application using familiar interfaces such as ODBC and JDBC. ADG implements ACID transactions at the cache level and supports full ANSI-SQL 99, including DDL and DML. Using Arenadata Grid, you can easily speed up transactional and analytical queries.

Arenadata Grid
ADG supports multiple languages and protocols

As part of the universal data platform, the Arenadata EDP, ADG can be integrated with the massive-parallel database Arenadata DB (based on the open-source Greenplum project), which allows multi-level data processing, benefiting from high-speed transaction processing in-memory and massive-parallel processing of a very large amount of historical data. Integration is performed via the PXF protocol.

Among the main use cases for ADG are the following:

  • speeding up the processing of analytical requests over operational data in a massive-parallel DBMS;
  • caching of operational data in HDFS;
  • implementation of transactional data cache for streaming systems and data buses.

Arenadata Grid
ADG can be integrated with Arenadata Hadoop and Arenadata DB using the HAWQ PXF service