March 2017 Summaries

At SingleStore, every company should have the opportunity to become a real-time enterprise. Data drives business, and data management performance is crucial for successful business results. The need for performance in today's 24×7 business world requires companies to ingest data quickly, execute low latency queries, and support a large number of analytics users and data volume simultaneously. Distributed systems can remove single server performance and capacity constraints, enabling the strength of many to act as one. SQL is seen as a prerequisite for analytics success, but SingleStore also supports other data models like JSON, key-value models, geospatial data, and connectivity to Spark. The flexibility to store data from memory to disk is also essential, and companies need cloud choice to deploy solutions on any server configuration or public/private cloud. With the right architecture, transactions and analytics can occur within the same system, eliminating the ETL process and consolidating database and data warehouses for immediate insight into critical applications.

I’m excited to be here to talk about machines and the magic of fast learning. We’re working with Thorn, a nonprofit dedicated to eradicating child exploitation online. They have to sift through millions of images daily using facial recognition and machine learning. With SingleStore, they can accelerate their image recognition workflow, allowing analysts to find matches faster and improve law enforcement response times. This has led to a 100-fold improvement in processing time, from 20 minutes to just 200 milliseconds. We believe that this technology can be applied to other areas with image-based datasets, such as mapping and social imagery, to gain insights through machine learning at scale. Our goal is to help businesses and nonprofits accelerate their operations with real-time data analytics.

This post by Mansour Raad of Esri discusses the integration of ArcGIS with SingleStore, a database that natively supports geospatial types. The integration uses the Spark Connector to ETL geo-spatial data into SingleStore in a Docker container, and extends ArcGIS Pro with an ArcPy toolbox to query SingleStore, aggregate and view results on a map. A comprehensive guide is available for learning more about SingleStore and its Geospatial functions, including design, build, and deployment of Spark applications using the SingleStore Spark Connector. Additionally, users can try SingleStore for free and explore its capabilities with a 79-page guide that includes code samples and performance recommendations.

The text discusses Eric Boutin's collaboration with Thorn, a non-profit organization that uses technology to combat predatory behavior and protect vulnerable children. Eric joined an engineering team at SingleStore, where he worked on optimizing the company's database for face recognition and machine learning tasks. The main challenge was matching fingerprints of images in real-time across a vast dataset, which Eric solved by adding linear algebra operators to SingleStore's architecture. This innovation enables Thorn to protect children more effectively and has implications for other use cases, such as click-through rate prediction and business intelligence analytics. Eric envisions future innovations that integrate machine learning frameworks with databases using SQL, enabling more complex models to be expressed in a scalable and efficient manner.

This is Strata+Hadoop World, a conference focused on data engineering and business intelligence, taking place in San Jose next week from March 14-16. SingleStore will be showcasing its capabilities to operationalize machine learning using real-time data, with keynotes and sessions on topics such as machine learning, predictive analytics, and Apache Kafka deployments. The company's CEO Eric Frenkiel will deliver a keynote on "Machines and the Magic of Fast Learning", highlighting advanced tools including TensorFlow, Spark, and Kafka, as well as compelling use cases demonstrating the power of machine learning to effect positive change.

The data warehouse as we know it has changed due to growth in data size and complexity, migration to the cloud, and the rise of real-time use cases, forcing enterprise organizations to expect more from their data warehouse. Gartner's latest Magic Quadrant Report has dropped data warehouse from its title and graduated to Data Management Solutions for Analytics, expanding the types of vendors covered and making inclusion criteria significantly more challenging. Real-time data has emerged as a top priority for data and analytics leaders, with high availability master data being a key capability as data volume, velocity, and variety increases. Multidomain and multivector MDM enable digital business to advance from collecting data for a single domain to connecting several domains, while improved packaged industry MDM solutions accelerate time to value and enable successful MDM programs. SingleStore has been positioned as a Challenger in the Magic Quadrant, citing broad use-case support, flexible deployment options, real-time processing, and customer engagement as factors leading to its movement to the top half of the quadrant.

The Gartner Data and Analytics Summit is taking place in Grapevine, TX, where SingleStore CTO Nikita Shamgunov will showcase the power of real-time data for Chief Data Officers transitioning to cloud-based applications. The summit features 8 tracks addressing top challenges facing data leaders, with sessions on the benefits of faster data warehouses, IoT pipelines, and machine learning. Keynote speakers include Gary Orenstein, Andrew White, Mark A. Beyer, Nikita Shamgunov, Roy Schulte, Alexander Linden, and others. SingleStore will also be exhibiting at booth 423, offering a free book on predictive analytics and machine learning, live demos, and prizes. The event coincides with Gartner's announcement that SingleStore has been positioned as a Challenger in the Magic Quadrant for Data Management Solutions for Analytics.