May 11, 2016 | Written by: Rod Hamlin
Categorized: Power servers | Power Systems
Share this post:
The next generation of enterprise applications is under increasing pressure to deliver insight at the speed of business. Interacting in real time with the customer requires more than historical intelligence. Applications must also anticipate a customer’s intent in real time. Data stores that serve such innovative applications need to offer high performance, deliver incredible versatility and be compatible with the enterprise needs of every environment.
Redis, the in-memory data structure store, has emerged as the next-generation database of choice. Redis is capable of delivering millions of operations per second with sub-millisecond latency on very little hardware, and it is remarkably proficient at handling multiple use cases–both operational and analytical, simultaneously.
To learn more about Power Systems, visit our website. And to learn more about how you can maximize performance with Redis and Power Systems, read the guide.
For example, look at Moovit, the world’s #1 transit app. It caters to 35 million users and handles 100 million daily passenger reports, 4.6 million bus and train stops and 4,500 transit operators. This is a perfect example of the challenges businesses face as they transform to meet the requirements of a digital world. The need to provide accurate information to commuters in real time requires that their infrastructure respond in less than 10 milliseconds. Failure means delays and frustration for consumers who rely on the service. To guarantee responsiveness and high availability, Moovit uses Redis Labs’ elastic, highly scalable database, to ensure customers get the insights they need when they need it. Redis provides not only the data ingestion and processing capabilities. Its data structures also facilitate real-time analytics with blazing fast speeds.
Another very common use of Redis is as a data-ingest engine or as a high speed cache front-ending other RDBMS and NoSQL databases. Its very high performance on very little hardware makes it an ideal choice when disk-based databases cannot handle millions of inserts and updates per second. Cooladata, a next-generation behavioral analytics solution, uses Redis Labs for real-time session-ization and analysis of user events (such as clicks and rollover,) across customer websites, while using MongoDB and Apache HBase for longer-term data storage and trend analysis.
When combined with IBM Power Systems, Redis Labs drives growth scenarios with low total cost of ownership (TCO) as you gain the additional horsepower that comes with having up to 8 threads per core (x86 systems have two). With the Redis Labs Enterprise Cluster (RLEC), you can seamlessly scale Redis to numerous instances, each handled by a thread.
RLEC and POWER deliver 200 percent more virtual processors or threads compared to Intel servers in the same space. The IBM Power Systems S822LC comes with 160 virtual cores, which means 160 instances of Redis can run on this box (more than 200 percent higher compared to Intel servers equipped with 48 virtual cores).
The resulting maximum throughput that can be achieved is 67 percent higher, with a 55 percent lower cost-per-transaction.
As the world expects increasingly faster processing and analysis of data, in-memory data structure stores like Redis which run 90 percent of the next generation of applications are increasingly gaining prominence. As more processing and analytics shift to Redis, it is worthwhile to consider choosing the more optimal hardware for such scenarios. Power Systems, with its high-compute capacity, innovations like CAPI-attached Flash and price performance leadership, is an ideal choice for enterprise-class Redis deployments.
Watch the webcast to learn how you can cut costs by up to 70 percent with Redis Labs and POWER8.
And click here to read the Spanish translation of this post.