OLTP

OLTP database

Scale-out OLTP Database

A MySQL-Compatible, OLTP Database with Lower TCO and Faster Time-to-Market

Clustrix provides the leading scale-out relational OLTP database engineered for the cloud or data center. ClustrixDB is a drop-in MySQL replacement built for high-transaction, high-value OLTP applications typically found in businesses such as ad tech, e-commerce, gaming, and large web and mobile businesses.

If your OLTP MySQL application is preventing your business from growing due to its scalability limits — there is only so far scaling up MySQL can take you — then you have to start investigating sharding and that means planning application and data architecture redesigns and development. ClustrixDB's distributed architecture, MySQL compatibility and linear, horizontal scaling can get you there faster with a lower TCO.

OLTP Database Overview

An OLTP database is different than data stores like Hadoop or Redis. OLTP databases need ACID guarantees, where you can always get the correct, most recent data. For online transactional processing applications, there is no margin for error to assume the data will eventually catch up. True OLTP databases are also relational, giving developers the flexibility of using joins in their queries. For developers looking for true relational databases with familiar SQL standards and web-scale strength, databases like MySQL have cemented themselves to be the cornerstone of most high-transaction, high-value applications.

Cloud-scale and OLTP

But cloud-scale is a completely different story, and one where most traditional OLTP databases fall apart. As the cloud has matured, developers increasingly look for cloud-native ways to power their data. They are looking for ways to elastically scale out their data using standard cloud instances for hardware. This has given rise to the data store, and its shared-nothing architecture that lets it take full advantage of distributed computing. Cloud database developers are also looking for portability across AWS, Azure and Google Cloud, as well as private clouds. Cloud vendors are starting to build databases that operate like this, like Google Spanner and Amazon Aurora, yet they are still relatively new and limited, and are bound to their infrastructure alone, failing the cloud-agnostic test.

A Cloud-ready OLTP Database

Ten years ago, the engineers at Clustrix set out to build a drop-in MySQL replacement that scales-out on commodity hardware, that is seamless for developers used to MySQL, and self-managing for admins. When we started, our choice to build an OLTP database that met these requirements was thought to be too hard. Of course, it was hard. We were conquering new ground, preserving what we liked from MySQL but building it from the ground up with a shared-nothing architecture, and making it easy to use. This took a lot of engineering to make an ACID-compliant SQL database that is truly distributed. Today, with the devops movement underway, the coin has flipped. Now sharding MySQL is too hard, constantly having to rebalance loads and refactor apps. Today limits on scale are unacceptable. So developers have made choices: trade ACID compliance for cloud-scale or face mountains of maintenance.

Clustrix is different. With ClustrixDB, developers do not have to choose.

The Database Landscape: OLTP Database versus OLAP Database

Databases can be confusing. While “relational” seems to be the type most commonly referred to when talking about databases, there are actually many other kinds of databases. In fact, there are hundreds of them available in the marketplace–not just relational–designed to handle different data types and manage a variety of workloads or use cases.

OLTP and OLAP

While there can be many subcategories of databases, most can be classified under two categories: transactional (or On-Line Transactional Processing) or analytical (On-Line Analytical Processing) databases.  OLTP databases are usually relational databases, and OLAP databases can be further subdivided into whether they are real-time analytics or data warehousing databases.

The OLTP Database

OLTP is characterized by a large number of short on-line transactions (INSERT, UPDATE, DELETE). The main emphasis for OLTP systems is put on very fast query processing while maintaining data integrity in multi-access environments; and effectiveness is measured by number of transactions per second at low latency and ability to handle high concurrency. There is detailed and current data in an OLTP database.

The OLAP Database

OLAP is characterized by relatively low volume of transactions. Queries are often very complex and involve aggregations. For these systems, response time is an effectiveness measure. There is usually aggregated, historical data in these databases.  Data warehousing usually looks at long historical data to find patterns and draw conclusions or recommendations about the present or future, while real-time analytics looks at transactions and/or data as they are happening in the moment– to draw similar conclusions or recommendations.

Built for OLTP. Not Your Average General-Purpose Database.

ClustrixDB is a distributed (multi-node), relational scale-out database designed for OLTP applications with high-value, high-transaction workloads deployed on-premises or in the cloud.  ClustrixDB is a perfect fit when you’ve outgrown your single-node OLTP database like MySQL, and need to scale beyond its limitations to handle higher transaction volumes and a massive number of concurrent users–without losing performance or ACID compliance.

Need more information?

Check out our resources page to view white papers, case studies, and more or get in touch with our experts.

"ClustrixDB is a holy grail of sorts in database solutions. Its ability to scale out was something we couldn’t achieve with MySQL without enormous expense and technical challenges.”

– David Smith, CTO, Viverae

viverae clustrix customer

"With ClustrixDB, we have not run into scaling issues anymore. As we need capacity, we just add nodes and see linear growth. We have seamlessly gone from a 3-node cluster to a 21-node cluster.”

–Nicolas Van Eenaeme, CIO of Twoo’s parent company Massive Media

twoo clustrix customer

"Over the last year we’ve seen significantly better performance and our December service incidents were reduced by 89% year-on-year.”

- Graham Hobson, CTO, Photobox

photobox clustrix customer

Ready to get started?

Contact us to request a free trial of ClustrixDB.