On SingleStoreDB v8.0. Q&A with Shireesh Thota
- You just announced v8.0 of your platform. What is special about it?
The need for delivering data-driven insights and for companies to respond in real-time is critical for customers to compete in today’s on-demand service economy. Data in this era is fast, accessible, and immediate – known as modern data – and the architecture needed to support this type of data requires real-time applications that allow for scale and faster and more dependable connections to support mission-critical workloads.
8.0 is the latest version of our platform that features faster analytics, improved developer experience, and greater ease of use. This update sets SingleStore apart as the only unified database for transacting and reasoning with data in real time in a multi-cloud hybrid distributed environment.
- Can you give us an example of how this new release makes it easier to use and how it enables faster developer adoption to power real-time and mission-critical applications?
SingleStoreDB’s new capabilities and improvements enable enterprises to protect their mission-critical applications and create faster analytics to easily derive value from all their data in real time. Customers like Siemens and Impact have realized real-time results since onboarding SingleStoreDB:
- Siemens: Before SingleStoreDB, Siemens Pulse Analytics Platform had limited capabilities due to how slow the system was. Data integration took 30 minutes. With SingleStore, Siemens saw a 10-100X speed improvement, which enabled real-time dashboards, and processes that formerly required ten steps were streamlined down to just one. With SingleStore, Siemens Pulse can support as few as 500 to as many as 100,000 concurrent users.
- Impact: Impact.com chose SingleStore over Snowflake, Cloudera, and BigQuery because of SingleStore’s ability to meet its growing data-intensive platform needs. Since SingleStoreDB, Impact.com has processed 20M events/ hour for 1,000 concurrent users. Impact has also seen a 1,000% reduction in query speed from 10 seconds to sub-second latency – something they couldn’t achieve with Cloudera Impala and Kudu.
About the latest release, Premal Shah, senior vice president of engineering and infrastructure and co-founder at 6sense says “SingleStore understands that business happens in real time” and offers “cutting-edge enhancements through its real-time, mission-critical applications that support our technology needs.”
Nuno Reis, IT director of architecture and transformation at Millennium bcp, states that with SingleStore, Millennium bcp is now able to “augment transactional applications with analytics, unlocking a real-time architecture by design.” This architecture is “key in building intelligent differentiated capabilities that drastically improve the customer experience.”
- What specific challenges of real-time and mission-critical applications are you addressing with this new release?
SingleStoreDB listened to its vast customer base and added many requested features that address the need for real-time and mission-critical applications.
Fast Seeking for JSON columns string data brings performance improvements up to 400 times faster than before. This is difficult in theory as data is stored in columnar format on a disk, which means it’s compressed. If you are trying to seek a particular set of data vs. do a fast scan of all the available data, you have to decompress a chunk which is usually in some kilobytes of data. Then you have to go through the metadata so that you don’t have to scan all the chunks, etc. Thus, making this process incredibly complex and time-consuming. Previously, SingleStore encoded JSON in a parquet format, so it didn’t support every transaction. In 8.0, we’ve redesigned the encoding completely so that Fast Seeking can directly access the middle segment of data without having to decode the entire thing.
Our Endpoint Load Balancer enhancements give our customers faster and more dependable connections to support mission-critical workloads. Previously, our Endpoint Load Balancer made it difficult to guarantee the high levels of stability that workloads demanded. We’ve created a new Endpoint Load Balancer that provides up to 30% improved load balancer connection performance, supports unlimited firewall rules and delivers increased connection stability. Customers like Impact and Armis have usability for a feature like this.
- What specific enhancement did you have in the area of analytics and query performance?
Advancements in analytics and query performance include:
- Wasm (WebAssembly) is now for both cloud or self-managed deployments, which opens the availability of this key feature to all users.
- Disk spilling is something you would see in a high-end OLAP system. It handles queries requiring more RAM than available. Previously disk spilling was available for hash group by operations alone (7.5, 7.8). Now disk spilling support is available for hash join, window and sort functions. All queries can now run as expected even if the resources required are more than available RAM.
- Tell us about the IBM Cognos connector. Why Cognos? What is it useful for and how does it work in practice?
IBM Cognos is an analytics software that has thousands of customers worldwide. However, to connect Cognos to SingleStoreDB, customers had to rely on MariaDB or MySQL connectors which were often unreliable, unsecured and resulted in connection errors. The problem was significant enough that some customers were willing to pay for third-party drivers (such as the one from Cdata.)
To avoid this, SingleStore partnered with IBM to build a native SingleStoreDB integration, or connector, for IBM Cognos Analytics.
The Congos connector eliminates these issues and provides a fast, reliable and secure connection between SingleStoreDB and Cognos Analytics to power analytical applications.
- What are Recursive CTEs? and what are they useful for?
Recursive CTEs is one our favorite updates for our developer community. CTEs without the recursion is a feature of SQL that allows for the definition of a temporary result set within a query. It helps simplify and maintain complex logic. And the recursive CTEs are a significant evolution, wherein the CTE can recursively invoke itself. Without this, querying hierarchical data like finding all employees under a manager, multiple levels down, for example, could be complex and cumbersome. If you bring in the power of recursion to CTEs, it makes it far more efficient. It simplifies a huge amount of logic into something simple, manageable, and easy to read without losing any performance issues necessarily. And this enables tree traversals and other graph operations in SingleStore.
Shireesh Thota is Senior Vice President of Products at SingleStore
Sponsored by SingleStore