Pavilion Data Resiliency and Availability Suite

Pavilion Data Systems, Inc. announced a suite of data resiliency features that enables customers to enjoy consistent performance and HA at the same time.

The company’s suite of performance and data resiliency features is designed around the belief that customers should not have to choose between performance and availability. The suite of features includes solutions that provide controller redundancy without compromising performance, protect vs. drive errors, and provide data redundancy while reducing capacity cost.

The firm’s Data Assurance technology is a patent-pending solution which helps guarantee data integrity. As SSDs grow in capacity, there is an issue, particularly in high I/O environments, where the SSD will write to the metadata that a write operation was performed, and yet fail to actually write the data. It is a rare, though significant error. The company solves this problem by writing a version number of the data to every 4k block, in addition to the standard CRC that is done as part of a T10 Data Integrity Field (Dif). By adding to existing data integrity protections, the solution eliminates the statistically significant number of errors that a standard T10 Dif does not.

The N+1 controller redundancy eliminates the need for customers to choose between performance and availability. Legacy dual controller arrays require the customer to choose between limiting performance to 50% of total capacity, by either running in active-passive mode or by reducing performance of each controller to 50% in active-active mode, or to run the risk of catastrophic performance degradation in the event of a controller failure. Capable of running up to 20 controllers in a single array, N+1 controller redundancy enables customers to enjoy the performance of multiple controllers running in parallel, with additional controllers operating in standby mode.

To protect vs. drive failures, the company’s Swarm Rebuild provides fast recovery in the event of a drive failure. Traditional drive protection takes the form of either erasure coding, which has a high capacity cost, or RAID, which can have unacceptably long rebuild times. Swarm Rebuild combines the low capacity overhead and dual parity data protection of RAID-6, with the ability to have multiple controllers Swarm a failed disk to rebuild arrays at the rate of 15mn per terabyte.

“Applications no longer have to compromise between maximum performance and HA because of Pavilion’s innovative architecture,” said Marc Staimer, president, Dragon Slayer Consulting. “Pavilion delivers unrivaled latency, IO/s, and throughput performance plus incomparable availability in one platform, delivering unsurpassed performance density.“

“As big and fast data applications become business critical, Pavilion allows customers to manage their modern workloads with the same SLAs as traditional workloads with its new features,” said V.R. Satish, co-founder and CTO, Pavilion. “Customers can now treat their modern workloads as tier-1 applications when it comes to data resiliency and availability.“

Read also:
Addition of Fast Object Storage to Pavilion Hyperparallel Flash Array Enables Customers to Get High Performance for Native S3 Data
Offering high performance block, file, and object storage on single unified platform [with our comments]
June 5, 2020 | Press Release
Pavilion Hyperparallel Flash Array Systems Certification With vSphere 7.0
Demonstrating 400% greater speeds and VM density using NVMe-oF
April 24, 2020 | Press Release