April 2017 Summaries

Ted Knudsen, CTO at Smartwaiver, uses Redis Enterprise in various ways throughout its application infrastructure, including powering kiosk applications with autofill functionality and maintaining daily statistics and dashboards. The database's data structures, such as Lists, Hashes, Sets, and Sorted Sets, are particularly useful for implementing complex functionality with simplicity. For analytics, Smartwaiver relies on Redis to run internal dashboards, leveraging data structures like Lists and Hashes to display recent waivers and more complex graphs. Caching is also a key use case, with Redis used to cache information about kiosks and implement webhooks. Compared to other in-memory databases like ElastiCache, Redis Enterprise Pack has proven to be a cost-efficient and high-performance choice for Smartwaiver's VPC environment.

The Redis developer relations team participated in the third HSHacks Hackathon, where teams from the San Francisco Bay Area came together to develop applications over a 24-hour period. The team provided support and guidance to participants, including a workshop on Redis, and sponsored a best Redis hack contest. One team, Vitruvian from Cupertino High School, won both the contest and the competition for using a healthcare dataset, developing an app that centralizes patient records to improve healthcare in rural areas and developing countries. The winning team utilized Redis's pub/sub functionality to process SMS messages received by the app.

The new docker image for Redis Enterprise Pack has been released as a preview version. This image extends the open-source Redis and provides high performance, linear scaling, and high availability with significant operational savings. It is available on Docker Hub and can be used to run Redis Enterprise Pack containers in various operating systems including MacOS, Linux, and Windows. The setup process involves running the container, setting up a cluster, creating a database, and connecting to it using redis-cli. Each container instance can run multiple open source Redis shards for seamless scaling. Additionally, Redis Enterprise Pack comes with a Proxy, Cluster Manager, and a simple visual UI for administration, alerting, and monitoring.

This approach automates building, storing, and deploying predictive models using a Remote Machine Learning Data Store hosted on Redis, leveraging DevOps CI/CD artifact pipelines to manage machine learning model artifacts across a team, allowing for efficient management of model deployments in intelligent environments. The workflow utilizes an API that can scale out expensive tasks, manages machine learning models with Redis caching, and natively archives models in S3, enabling organizations to focus on improving model predictive accuracy, dataset features, and deploying pre-trained models to new environments for automation and real-time predictions.

We've seen significant performance improvements of up to 2.6 times with Redis Enterprise Flash (Redise Flash) on AWS I3 instances compared to I2 instances, which are based on SATA SSDs. The new I3 instances offer much lower latency and higher throughput thanks to NVMe technology, resulting in substantial cost savings. With Redise Flash, we can store frequently accessed data in RAM and less frequently used data on flash storage, making it an attractive solution for businesses with large datasets that require high performance and low costs.