Solving your performance problems with Intel Optane
Organizations often face problems with their current database infrastructure regarding performance, data persistence, and cost. Some may use in-memory databases to help address performance concerns but they may run into data persistence issues. Sacrificing the integrity of your data is never a good tradeoff for performance. Also, more memory in a server will drive up costs. An ideal solution would provide the performance of an in-memory database without compromises on throughput, latency, and data persistence.
Luckily, ScyllaDB and Intel teamed up to solve these problems so organizations can get the most out of their hardware by using ScyllaDB with Intel Optane SSD’s to resolve cold-cache and data persistence challenges. If you are familiar with ScyllaDB, you already experienced optimal performance and lower latencies compared to other databases. By adding Intel Optane drives to the mix, ScyllaDB users achieve breakthrough performance and even lower costs.
At ScyllaDB Summit 2017, Frank Ober of Intel and Tomer Sandler of ScyllaDB walked through the performance testing in depth. You can learn about Intel Optane performance this blog post.
When using Optane drives, the cost of a typical ScyllaDB configuration will drop significantly due to the reduced number of DRAM required per server. Looking at the image below, we can see that a typical in-memory solution will cost more than double to achieve the performance of a combined ScyllaDB and Intel Optane solution.
Together, ScyllaDB and Intel Optane Drives achieve the performance of an in-memory database, and resolve cold-cache and data challenges without compromising on throughput, latency, and performance, while delivering a lower total cost of ownership.
Here is a summary of the benefits using ScyllaDB and Intel Optane drives.
To learn more, you can view the entire presentation below or watch the video above:
Want to test ScyllaDB’s performance yourself? You can take ScyllaDB for a test drive for free. Our Test Drive lets you quickly spin-up a running cluster of ScyllaDB to see for yourself how it performs.
