Wide Range of Applications

Stateful stream-processing with digital twins offers deeper introspection than previously possible across a wide range of applications, including financial services, e-commerce, Internet of Things (IoT), medical systems, manufacturing, logistics, and much more.

Here’s just one example: A medical monitoring system can use a dynamic, digital twin model of a patient combined with machine-learning techniques to take into account a patient’s detailed history, medications, and current status when evaluating incoming device telemetry. This allows the system to generate more intelligent and timely alerts to medical professionals.

Breakthrough Architecture

Unlike mainstream platforms such as Apache Flink, Spark, and Storm, ScaleOut StreamServer automatically dispatches incoming streaming events to in-memory, object-oriented, digital twin models of data sources for analysis. Other platforms need to pull state information from remote data stores, such as database servers and distributed caches; this creates delays and network bottlenecks. ScaleOut StreamServer processes incoming data streams within digital twins stored in an in-memory data grid — where the data lives — ensuring minimum access latency and peak stream-processing throughput.

In-Memory, Data-Parallel Computing

ScaleOut ComputeServer brings the power of data-parallel computing to distributed, in-memory data. Its compute engine minimizes data motion and delivers blazingly fast execution times. ComputeServer uses techniques from parallel supercomputing, such as multicast and global combining, to deliver results as fast as possible. These capabilities are integrated into an intuitive, easy to use SDK with automatic code shipping that makes application development in Java, C#, and C/C++ simple and straightforward.

Unique Capabilities for Operational Intelligence

Unlike in-memory computing platforms that focus on analyzing streams of data, ScaleOut ComputeServer was specifically designed for operational intelligence within live systems. Its object-oriented, in-memory data storage lets you create a scalable, highly available model of real-world entities (such as portfolios, e-commerce shoppers, or the Internet of Things) which easily can be updated as changes occur. Data-parallel computing enables fast analysis to detect emerging patterns and generate immediate feedback. You have a mission-critical system, and now you have the operational intelligence you need for peak performance.

Integrated Scalability and High Availability

Scalability is in our DNA. Every aspect of our in-memory data grid’s architecture is designed to deliver linearly scalable performance with integrated high availability. This means that you can add servers to meet the demands of your business and always expect fast response times in accessing your data. If and when server failures occur, our patented technology ensures that your data will be there and your data grid will recover fast and reliably.

Ease of Use to Keep Your TCO Low

Whether it’s deploying and managing a ScaleOut cluster or developing a new application, we make ease of use a top priority for our products. Our goal is to take care of the grid’s inner workings and keep the APIs as simple as possible so that you can focus on your application’s structure and performance. ScaleOut StateServer automates cluster membership, data partitioning, and load-balancing, and it self-heals with automatic rebalancing after a server failure.

In-Memory MapReduce for Real-Time Results

You’ve made an investment in MapReduce and Hive, but with traditional Hadoop you can only realize that investment in batch processing on static data sets. With ScaleOut hServer, you can run MapReduce or Hive on both HDFS data and fast-changing, in-memory data. Sub-second scheduling time and numerous optimizations for real-time execution (such as optional sorting) ensure fast results. Benchmark results have demonstrated 40X speedup and faster.

Fast Development and No Tuning Required

Installing Hadoop and tuning applications for high performance can be challenging. Sidestep the complexity with ScaleOut hServer. Install it on a cluster of commodity servers – or even your laptop – and run your first MapReduce application in minutes. Self-tuning for in-memory data ensures peak performance. ScaleOut hServer is compatible with the most popular Hadoop distributions, including Apache, Cloudera, and Hortonworks. Use Hadoop YARN to run both in-memory MapReduce/Hive and standard batch processing in one, integrated deployment.

Rock Solid Disaster Recovery that Eliminates Service Interruptions

With ScaleOut GeoServer you can automatically replicate data as it changes among up to eight data centers to protect against data center-wide outages. It’s easy to set up and manage, and it supports “master-master” usage for maximum cost-effectiveness. ScaleOut GeoServer delivers scalable throughput that automatically grows as you add servers to each data center, and it automatically handles server failures. This product has been battle-tested by some of the world’s leading e-commerce sites.

Unified Access to Data in Different Area Codes

ScaleOut GeoServer can combine your in-memory data grids at multiple data centers, both on-premise and cloud-based, into a single, logical data grid. We call this global data integration, and it dramatically simplifies your application, letting it transparently access data when needed from any location. It even supports distributed locking across sites so that you can globally synchronize updates.