Data-intensive applications require speed and scalability to facilitate real-time decision-making.
The 42nd IT Press Tour had the opportunity to meet with Suresh Sathyamurthy, CMO and Domenic Ravita, VP of Product Marketing at SingleStore, the cloud-native database built with the speed and scale necessary to power data-intensive applications.
Businesses are becoming more data-intensive. Data volume and complexity are rising. Delivering the best customer experience (CX) requires real-time analytics. Data-hungry AI/ML models are being applied to predict outcomes and recommend products.
Applications are struggling to keep up. The sluggish event-to-insight response is unable to meet SLAs. There are rising costs and complexity. Growing user demands drive increased concurrency challenges.
Data-intensive applications need the following attributes: 1) data size with the ability to handle hundreds of terabytes or petabytes of data; 2) speed of ingestion with thousands to millions of rows/second; 3) latency requirements of sub-second to millisecond latencies; 4) complexity with a high number of joins in the queries; and, 5) concurrency with tens to hundreds of users accessing the application simultaneously.
Cloud Data Evolution
The first era of the cloud data evolution from 1999 to 2020 saw a retrofitting of on=premise and general-purpose SQL databases. The second era from 2010 to 2017 saw a retreat from SQL to achieve speed and scale for the adoption of special-purpose NoSQL databases and the adoption of cheap object storage. The third era of the cloud data evolution began in 2017 with real-time, hybrid multi-cloud, multi-model, and relational databases.
The third era of cloud data evolution has driven SingleStore to develop a frictionless and ultrafast, distributed SQL, multi-model, hybrid multi-cloud, databases for real-time applications.
Several trends are driving database market disruption. The internet is getting faster fueling hyper-growth in the adoption of modern applications for streaming, gaming, IoT, and other interactive real-time applications. Modern applications are driving the need for converged streaming data, transactional, and analytical processing on multi-model modern data. Applications are being re-platformed and the world is going hybrid multi-cloud.
The Future
SingleStore believes companies will consolidate their database technologies. The future will be hybrid multi-cloud, frictionless, and ultrafast for real-time decision-making. The convergence of OLAP and OLTP will power modern applications. Distributed SQL is here to stay.
The popularity of SaaS applications continues to grow as both customer-facing and internal applications are powering businesses. Customer experience is driven by immersive customer-facing apps while business experience is being driven by applications with millisecond latencies to handle operational workloads.
The requirements for the data tier of modern SaaS applications are operational (real-time update/delete/transaction), fast analytics (leverage RAM + CPU parallelism), highly scalable (sharding), high availability (replication), SQL for BI access, and start free and grow to global scale.
SingleStore DB
According to Jordan Tigani, Chief Product Officer at SingleStore, "People who are just learning about SingleStore often don't believe that transactions and analytics can be unified into a single database."
The advantages include scalable, relational SQL from distributed-native, bottomless, cloud-native, ACID, ANSI SQL. Unified single store, document, key-value, time-series, geospatial, and streaming data. Optimized for real-time with parallel ingestions, vectorization, lock-free, MVCC, SIMD, NVMe compressed storage, and efficient distributed joins.
The key capabilities are:
Universal storage -- patented single table type for transactions and analytics. The only database with a single table type for both.
Limitless point-in-time recovery.
Pure speed -- fast transactions, analytics, vectorization, and query compilation.
Fast ingestion -- pipeline load with data updates.
Unlimited storage -- separation of storage and compute in an operational analytic database.
Multi-model -- relational foundation with first-class support for document, key-value, geospatial, time series, and full-text search.
Run anywhere -- hybrid multi-cloud, SaaS, on-premises, and Kubernetes operator.
Compatibility -- ANSI SQL, MySQL & MariaDB ecosystem.
Data Intensity Assessment Tool
SingleStore is developing a quick assessment tool that will be available on their website to determine how data-intensive an environment is based on five variables: data size, query latency, complexity, data ingest speed and concurrency.
The tool is for senior developers and engineers, application architects, database architects, DBAs, and data engineers. Developers answer a series of questions about the five characteristics of their workload and responses are scored in a document they can share. The tool is being built to help educate developers about data intensity and help them identify their data-intensive workloads and applications.