June 2022 Summaries

We are restarting developer meetups! Join us on July 19th in our Mountain View office to learn from experts about robust monitoring of Aerospike nodes and AI/ML-based real-time fraud detection systems, as well as new releases of Aerospike Monitoring version 2.0.0 and Prometheus exporter version 1.6.0. The Firefly Alpha Program for Aerospike's graph database offering has been announced, along with a blog post on data modeling for speed at scale. Community questions have also been answered regarding cluster configuration and node performance.

Aerospike Database is a purpose-built database for applications requiring predictable sub-millisecond access to billions and trillions of objects, storing many terabytes and petabytes of data while maintaining small cluster sizes. Collection Data Types (CDTs) are essential for modeling related objects in one record, allowing transactional semantics across multi-object updates. CDTs include List and Map data types, which can be nested at an arbitrary level, providing a compact tuple form for storing objects as Lists, eliminating bin overhead, and offering direct access to elements by value if the id is stored as the first field. Aerospike's CDTs support performance aspects such as multi-element updates, single, multi, and range selectors, and multiple return types, enabling efficient server-side execution and minimizing data transfer cost. The use of CDTs as containers unlocks advantages toward speed-at-scale, including high-density storage, fast access to related objects, and efficient querying mechanisms like filter expressions and secondary indexes. CDTs are integrated into Aerospike's performance features, including sets, set indexes, multi-op requests, batch requests, transactions, and consistency, making it a suitable choice for applications requiring real-time data processing and scalability.

Relational databases store data in rows and tables with links between information, while non-relational databases use a storage model tailored to the type of data being stored. Relational databases are known for their simplicity, consistency, accuracy, reliability, flexibility, normalization, and security. Non-relational databases excel in low latency, availability, flexibility, simplicity, less maintenance, and scalability. When choosing between the two, consider factors such as data type, resources, real-time data needs, scalability, and performance demands.

Aerospike, a data processing platform, is addressing scale and performance needs for IoT data in utilities through low-latency data processing, strategic partnerships, and the addition of time-series API. The U.S. power grid is aging and strained, with demand for more power during extreme weather fluctuations due to global warming. IoT utility market spending is expected to grow from $28.6 billion in 2019 to $53.8 billion by 2024. Aerospike's data platform can complete the entire cycle of ingesting and indexing/partitioning data in less than 30 minutes, compared to three hours for some utility companies. The platform is also well-suited for mixed workloads and can be leveraged for other use cases such as analytics that require high read throughput. Aerospike's ability to power IoT ecosystems includes partnerships with Ably, an edge messaging platform designed with resilience at its heart. The combined solution is a great fit for both enterprise-grade and mass-market IoT use cases that need data ingestion and processing at speed and scale with resilience and guaranteed performance. Aerospike's time series API leverages bespoke features to provide efficient reading/writing of time series data at scale, making it capable of supporting a wider range of use cases.

Amazon Web Services (AWS) has introduced I4i instances based on the 3rd Generation of Intel® Xeon® Scalable processors, offering superior performance for Aerospike compared to I3/I3en instances. The I4i instances provide greater turbo performance and up to 2x more DRAM than I3 and I3n instances while supporting storage-dense configurations of up to 30TB. In benchmark tests, the I4i.32xl instances achieved 18% higher single-SSD throughput compared with i3en.24xl instances. The availability of AWS I4i instances allows for cost-effective configurations at scale across a wider range of applications.

Aerospike has tested I4i instances based on Intel Xeon Scalable 3rd Generation processors and found them to be an attractive alternative to I3/I3en instances for many workloads, particularly those with high storage and CPU requirements. I4i instances offer greater turbo performance and more DRAM than I3 and I3n instances, while supporting storage-dense configurations of up to 30TB. Aerospike has compared the performance of I4i.32xlarge against i3en.24xlarge instances in various tests, including benchmarking using Aerospike's Certification Test program and testing 3-node clusters with client instances running the Aerospike Java Benchmark tool. The results show that I4i instances achieved higher single-SSD throughput, improved linearity of SSD performance, and lower latency compared to i3en instances. This suggests that I4i is a suitable choice for applications requiring high CPU cores or DRAM, while supporting high-density SSD storage and network bandwidth.

Aerospike has introduced Aerospike SQL Powered by Starburst, a solution that significantly reduces the time needed for business insights and real-time decision making. This new offering allows customers to run complex SQL queries on data stored in the Aerospike Database at gigabyte-to-petabyte scale. The integration of Starburst's SQL engine with Aerospike provides a complete SQL solution, opening up the platform for more analytics use cases and enhancing revenue streams for partners who can resell this OEM solution.

Aerospike SQL powered by Starburst enables businesses to generate valuable insights rapidly from their data stored in the Aerospike NoSQL database using ANSI SQL queries. This solution combines Aerospike's low latency and high throughput capabilities with Starburst Enterprise, a fully supported distribution of open-source Trino (formerly Presto SQL). The Aerospike SQL connector bridges the data model differences between NoSQL databases like Aerospike and SQL engines like Trino. This integration allows users to run complex queries, create insightful dashboards, and perform in-depth data analysis using familiar business intelligence tools such as Tableau and Power BI. With features like massive parallelism, predicate pushdown, and secondary indexes, Aerospike SQL boosts query performance and enables businesses to make faster, more informed decisions.

Aerospike Inc., a real-time data platform leader, has announced Aerospike SQL Powered by Starburst, a solution that enables customers to run massively parallel, complex SQL queries on petabyte-scale data stored in the Aerospike Real-time Data Platform. The new solution is built on top of the popular and low-cost open-source Trino platform and provides a single point of access to federated data using existing SQL analytic tools such as Tableau, Qlik, Power BI, and others. This enables data analysts and data scientists to dramatically reduce the time to insight on large volumes of real-time data and drive faster decisions for better business outcomes. The partnership between Aerospike and Starburst aims to democratize data for enterprise SQL developers, data architects, and BI analysts, allowing them to achieve fast, efficient SQL queries at an unlimited scale.

Account takeover fraud (ATF) is a growing problem, with an estimated 22% of American adults experiencing it. The trend has been driven by the expansion of online financial services, which are predicted to grow to $310 billion. ATF can lead to significant losses for victims and damage companies' reputations. Strategies to combat this issue include using machine learning at the edge to analyze large amounts of data quickly, improving AI/ML algorithms to reduce false positives, and real-time ingestion of customer, market, and transactional data from multiple sources. Aerospike's real-time data platform offers these capabilities while maintaining high reliability and low total cost of ownership.