If your organization has been avoiding flash-based storage because of concerns about losing mission-critical data in the event of a disaster, Axxana’s Phoenix System provides the ideal solution—without imposing the performance drag associated with synchronous replication.
Flash Storage = Flash Performance
As the cost of flash-based storage solutions declines, many organizations are considering flash arrays in order to take advantage of the significant performance benefits inherent in their design. Unlike traditional hard disk drives (HDDs), which rely on mechanical processes (think “booting up” and “spinning”) to read and write data, flash-based storage is built around solid state drives (SSDs), which rely on electrical processes. This design accommodates extremely high input/output (I/O) rates (i.e., throughput) and minimizes data access times (i.e., latency). It also ensures persistent data integrity when power is off, reduces power consumption, and more. Given its attributes, flash-based storage is ideal for virtual desktop infrastructures, business-critical applications that support online transaction processing, and other database operations with heavy workloads.
Synchronous Replication Wipes Out Performance Gains
What complicates matters is that business-critical applications typically store business-critical data, which must be protected from loss or damage in the event of a disaster. In a traditional disaster recovery model, an organization uses synchronous replication to ensure that essential data is not lost. However, synchronous replication is not feasible for flash-based storage for the following reasons:
- High data transmission costs. Synchronous replication must be executed across data transmission lines that support minimum latency in order to ensure zero data loss in case of a disaster, without delaying the application. Synchronous replication solutions require costly fiber optics or other dedicated cabling. Flash-based storage imposes such high I/O rates that it is almost impossible to find bandwidth that matches the extremely low latency requirements of flash-based storage arrays.
- Additional latency. With synchronous replication, each write operation must be propagated to a remote site. This propagation delay imposes additional latency on write performance, and it robs organizations of the performance gains that cause them to seek flash-based storage in the first place.
Axxana’s Solution: Zero Data Loss Without the Performance Overhead
The Axxana Phoenix System addresses these challenges by allowing organizations to replicate asynchronously to any distance, while still protecting data that has not yet been written to the remote disaster recovery site. In this scenario, organizations can ensure zero data loss for all their business-critical applications. That is, they can achieve recovery point objective (RPO) = 0 while meeting the following goals:
- Reduced data transmission costs. Organizations can maintain unreplicated data (i.e., lag) at the primary site. This lag is protected inside the Phoenix System’s disaster-proof “black box,” so in case of a disaster, it is not lost or damaged. Minimizing the amount of data to be transferred significantly reduces the bandwidth requirements to the remote site.
- Minimal latency. By eliminating synchronous replication—and the associated propagation delays—to the remote site, organizations can significantly reduce the latency overhead on write operations. Although synchronous writes occur to the Phoenix System, overall latency is minimal because the Phoenix System is in the same data center as the primary storage and/or databases.
Ready to get started with a flash-based storage solution that meets performance and disaster recovery goals? By using the Axxana Phoenix System, you gain the disaster-recovery benefits of synchronous replication, while affordably protecting all critical data across the full range of critical applications.
For more information firstname.lastname@example.org
For more on Axxana’s solution for flash-based storage in Oracle Data Guard environments, see Using Flash-Based Storage Without Compromising Transactions (or Performance).