It’s been about a year since we first introduced Swarm7, which was based on emergent behavior for the sole purpose of providing scale-out cloud storage that adapts to any use case or workload. Its evolution has continued, and today we’re announcing significant enhancements based on your feedback and real-world use cases.
Swarm is built according to the principles of emergent behavior—the creation of ordered, interdependent systems out of simple and nearly independent elements for a whole that is greater than the sum of its parts. Social insect colonies are the classic examples, such as ants and bees. At Caringo, we are firm believers in biomimicry, imitating the simple and elegant ways nature solves its problems as described in our whitepaper Emergent Behavior: The smarts of the Swarm. (We even patented a synchronization algorithm based on the way nature has worked for more than a billion years.)
In the latest version of Swarm, we have enhanced the architecture to extend the principles of emergent behavior to issues found in big data storage and archive environments:
Retiring and refreshing hardware more efficiently
When you upgrade to new hardware, add capacity, or retire failed drives, there is often a performance impact or a time-consuming sequence of steps involved. We’ve developed a leading-edge process to limit the impact of capacity balancing during hardware maintenance and refreshes—all without compromising data integrity. In addition, we developed a single-step, synchronous process to make data swaps more predictable and efficient. You can now limit operations to specific nodes, providing greater control during hardware maintenance operations.
For customers who routinely erase large amounts of data—such as cloud-based digital video recording (DVR) solutions—this feature also ensures capacity is reclaimed and available as quickly as possible. Customers performing large data transfers, such as moving old hardware out of a cluster, will also benefit from smarter load balancing.
Handling error conditions
Swarm has always proactively managed drive errors, but the latest version handles read errors differently than it handles write errors. We’ve added more intelligence in monitoring disk health to give you more insight. We’ve also added a widget to the console to monitor memory conditions for faster issue resolution—all while making sure that your data remains protected.
Future-proofing your storage
To help future-proof cloud storage solutions, we’ve added multipath support so you can take advantage of dual SATA connections and newer, denser storage arrays for high availability, high performance and seamless migration to web-scale capacities.
Making information easy to find
If you are an existing customer, you can find information about these enhancements in our updated, searchable product documentation found in Caringo Connect. Also, feel free to drop us a line at email@example.com.
Elasticsearch is a distributed, RESTful search and analytics engine that can be used with object storage to enhance searchability of metadata searching operations. Join Caringo VP of Product Tony Barbagallo and Caringo Engineer Jamshid Afshar … More Details »