Riak TS
Riak TS is an enterprise-grade NoSQL time series database from Basho Technologies, optimized specifically for IoT and time series data workloads.
Publisher review
Riak TS is an enterprise-grade NoSQL time series database from Basho Technologies, optimized specifically for IoT and time series data workloads. It ingests, transforms, stores, and analyzes massive amounts of time series data, and is engineered to be faster than Cassandra for these use cases. The database is designed for companies that need to make real-time decisions, reduce costs, increase performance, and manage large-scale data aggregation.
It uses a masterless architecture that automatically replicates data across nodes, ensuring high availability even during hardware or network failures. Riak TS supports scaling from one to over 100 nodes on commodity hardware with near-linear performance improvement, and it automatically distributes data uniformly across the cluster. Key capabilities include SQL commands for defining semi-structured data, SQL range queries for time-based queries, built-in aggregations for rollup and analysis, and data expiry for deleting unnecessary data.
Data co-location based on time range enables faster read access. In a demonstration, Riak TS loaded over 350,000 records from the Bay Area Bike Share open data set to illustrate its read and write speed. Compared to Cassandra, Riak TS is positioned as faster for time series workloads, with a simpler operational model that makes it easy to add and remove capacity on demand.
However, Riak TS has notable trade-offs: it lacks a free text search function, requires more security work, does not support multi-tenant reporting, and has limited types of index optimization. Additionally, automating repairs after unclean shutdowns is not supported, and WebDAV/Samba shares for Riak CS are not available. Implementing SQL queries from Riak TS in Riak KV is not supported, settable replication bandwidth caps are not available, and safemode startup after failure is not supported. Pricing is available only upon request, and the company is headquartered in Bellevue, Washington.
How it works
-
Masterless architecture
Automatically replicates data across nodes to ensure high availability even during hardware or network failures.
-
SQL range queries
Enables time-based queries using standard SQL select statements with a specified time range and chosen fields.
-
Built-in aggregations
Reduces query time dramatically when analyzing large data sets by providing rollup and analysis functions.
-
Data co-location
Co-locates data in the same physical part of the cluster based on time range for faster read access.
-
Data expiry
Allows automatic deletion of unnecessary data to manage storage and maintain performance.
-
Scalable on commodity hardware
Easily scales from one to over 100 nodes with near-linear performance improvement using commodity hardware.
-
ANSI-compliant SQL
Uses standard SQL for queries, making it accessible to a wide range of developers and business analysts.
Strengths and trade-offs
Strengths
- High resiliency with automatic data replication across nodes ensures data availability even during hardware or network failures.
- Scalable, masterless architecture allows adding or removing capacity seamlessly on commodity hardware with near-linear performance improvement.
- Operational simplicity makes it easy to set up, upgrade, and scale clusters up or down on demand.
- Fast reads and writes of time series data, demonstrated by loading over 350,000 records from the Bay Area Bike Share open data set.
Trade-offs
- Lack of free text search function limits query capabilities for non-structured text data.
- More security work is needed compared to some competitors, increasing operational overhead.
- Multi-tenant reporting is not available, making it unsuitable for shared environments with separate reporting needs.
- Limited types of index optimization restrict query performance tuning options for complex workloads.
Pricing context
Pricing available only upon request; no public tiers or dollar figures are disclosed.
Getting started with Riak TS
-
Download and install Riak TS
Download the Riak TS package from Basho's official site for your operating system. Follow the installation guide to set up the database on a single node or cluster, ensuring all dependencies are met.
-
Start the Riak TS cluster
Start the Riak TS service on each node using the provided command-line tools. Verify that all nodes join the cluster and that the masterless architecture replicates data automatically across them.
-
Define a time series table
Use SQL commands to create a table with a primary key that includes a timestamp and partition key. Specify columns for your data, such as sensor ID and value, to structure the time series data.
-
Load time series data
Insert sample time series records into the table using SQL INSERT statements. For bulk loading, use the provided client libraries or command-line tools to ingest data from files like CSV.
-
Run a time-based query
Execute a SQL SELECT query with a WHERE clause on the timestamp column to retrieve data within a specific time range. Use built-in aggregations like COUNT or AVG to analyze the results.
Frequently Asked Questions
What is Riak TS and what is it used for?
Riak TS is an enterprise-grade NoSQL time series database from Basho Technologies optimized for IoT and time series data. It ingests, transforms, stores, and analyzes massive amounts of time series data, helping companies make real-time decisions and manage large-scale data aggregation.
How does Riak TS handle high availability and data replication?
Riak TS uses a masterless architecture that automatically replicates data across nodes. This ensures high availability even during hardware or network failures, as data is uniformly distributed and accessible from any node in the cluster without a single point of failure.
What SQL features does Riak TS offer for querying time series data?
Riak TS supports ANSI-compliant SQL for defining semi-structured data and executing range queries based on time. It includes built-in aggregations for rollup and analysis, and data co-location by time range speeds up read access for time-based queries.
How does Riak TS compare to Cassandra for time series workloads?
Riak TS is engineered to be faster than Cassandra for time series workloads, with a simpler operational model for adding or removing capacity. In a demonstration, it loaded over 350,000 records quickly, highlighting its read and write speed advantages.
What are the main limitations of Riak TS?
Riak TS lacks free text search, requires more security work, and does not support multi-tenant reporting. It has limited index optimization options, no automated repair after unclean shutdowns, and no settable replication bandwidth caps, which can increase operational complexity.
How scalable is Riak TS on commodity hardware?
Riak TS scales from one to over 100 nodes on commodity hardware with near-linear performance improvement. Its masterless architecture automatically distributes data uniformly across the cluster, making it easy to add or remove capacity on demand without performance degradation.
Alternatives
How Riak TS compares
Direct head-to-head against 2 competitors. Picked by 7wData.
Riak TS
- Pricing
- Pricing available only upon request; no public tiers or dollar figures are disclosed.
- Target
- Riak TS is an enterprise-grade NoSQL time series database from Basho Technologies, optimized specifically for IoT and time series data workloads.
- Strength
- High resiliency with automatic data replication across nodes ensures data availability even during hardware or network failures.
- Watch for
- Lack of free text search function limits query capabilities for non-structured text data.
Redis
- Pricing
- Free tier available, enterprise plans custom/contact sales
- Target
- High-performance caching, real-time apps
- Deployment
- Cloud, on-prem, hybrid
- Strength
- Sub-millisecond latency, rich data types
- Watch for
- Memory-bound architecture, scaling costs
RavenDB
- Pricing
- Free community edition, enterprise $1,999/server/year
- Target
- ACID-compliant document store
- Deployment
- Cloud, on-prem, hybrid
- Strength
- Built-in full-text search, transactional support
- Watch for
- Smaller community vs. MongoDB
User reviews
No user reviews yet. Be the first to write one.
Sources
Reporting on this tool draws on these publicly available sources.