ESG's Mike Leone discusses the results of ESG Lab's recent testing of WekaIO Matrix, a software-defined storage solution.
Welcome to another ESG Lab Video Summary where I'll be reviewing the results from our recent testing of software-defined storage solution, WekaIO Matrix.
As data continues to grow, organizations are rapidly adopting software-defined storage solutions with a goal of simplifying the overall storage experience while meeting strict IT budget requirements. In fact, recent ESG research shows that the top reasons for evaluating software-defined storage are simplifying management and reducing costs.
We look to validate WekaIO Matrix's ability to deliver on these factors, focusing on deployment of management simplicity, data resiliency, and performance at scale. WekaIO Matrix is a software-defined storage solution. It's POSIX comply in virtual file system leverages off-the-shelf servers and SSDs to cost-effectively provide high performance, high capacity, and resilient storage at scale, both on premises and in the cloud on AWS.
We began the installation process by logging into the WekaIO GUI on one of the servers. We selected the installation type, and then went through the configuration wizard to complete the initial deployment, which took under eight minutes. After confirming all the entered information, we formed the new WekaIO cluster, which was I/O ready in under a minute.
For validation and management simplicity, we toured the GUI, served as a well-organized UI for managing and reporting everything on the WekaIO Matrix cluster. The "System Overview" tab displayed details about each node as they were selected, while the "Statistics" tab displayed a histogram of events in the cluster that could be drilled down on for more details. We could add any statistics that we thought were relevant, and by looking at event balloons, we could visually correlate cluster-wide events to specific data measurements.
Next, we switched to a 16-node WekaIO Matrix cluster running on AWS to create a local and tiered file system. In both cases, the same process was used, with us setting up the parameters of each option and simply clicking "Create." For each case, the whole process took about a minute.
For resiliency and data protection, we learned about MatrixDDP, which applies patented, error-correction algorithms to manage how data is distributed across a cluster. Local and cloud clustering improves I/O and rebuild performance since WekaIO maintains performance and availability as the cluster size grows.
In the event of multiple failures, MatrixDDP prioritizes rebuild starting with the data that is least protected to return the cluster to the next higher level of resiliency as soon as possible. We viewed a rebuild in process which demonstrated a 16-node cluster going from N+1 resiliency back to full N+2 resiliency in less than 15 seconds
Finally, we audit tests run by WekaIO to evaluate performance and scalability potential. For testing in the cloud, a cluster was sized at 30, 60, and 120 nodes in a hyper-converged configuration to highlight the predictable performance as the cluster grows. When evaluating IOPS, 4KB writes and reads yielded peak performance at the 120-node count, reaching 3.76 million IOPS for reads and 399,000 IOPS for writes, while response times for reads and writes remained under 1 milliseconds for all test cases.
We also analyzed throughput capabilities in the same cluster for 1 megabyle reads and writes. Again, we saw linear performance scalability as the test bed doubled in size, from 30 through 60 to 120 nodes. At the 120-node count, ESG witnessed 59.1 gigabytes/second throughput for reads while writes yielded an impressive 34.7 gigabytes/second.
The WekaIO software-defined storage solution enables deployment flexibility for hosts on premises or virtual clients in the cloud. Through the use of NVMe and SSD storage, large-scale capacity and performance is achieved that rivals that of a traditional, on-premises all-flash array, while advanced data protection with MatrixDDP enables high levels of protection and resiliency at massive scale. These capabilities are easily portable to the cloud with cloud-aware compute and storage that's able to expand and contract on demand based on the current needs of the business to allow organizations to keep costs in check.
To address the complexities that come with constantly growing data sets and their strict performance, recovery and cost SLAs, enterprises are rapidly adopting software-defined storage to replace traditional, complex, and costly storage solutions. ESG recommends WekaIO as a cost-optimized software-defined storage solution that combines cloud-like scale, agility, and protection with on-premises, all-flash storage performance.
If you'd like to learn more, read the full report.