AccelStor proudly offers a true high availability all-flash array (AFA) with outstanding reliability and fault tolerance, thanks to the concept of a shared-nothing architecture. The NeoSapphire High Availability (HA) model ensures no single point of failure, which means essentially zero downtime. Given today’s always-on public/private cloud applications, non-stop service is vital, especially for enterprise-level business. Server malfunctions can be caused by hardware failure, human error or even natural disasters — but with the NeoSapphire HA all-flash array, you can depend on consistent service with minimal risk of unexpected interruption and costly downtime.
The NeoSapphire High Availability AFA is built on a shared-nothing architecture — a distributed computing concept wherein multiple storage nodes are independent and self-sufficient. Each node has its own set of resources, including CPU, memory, network and dedicated storage pool. These nodes execute synchronous replication with each other via a high-speed, low-latency InfiniBand FDR (fourteen data rate) interlink. This eliminates the risk of failure caused by shared components. This approach, specifically designed to provide no single point of failure and minimal service interruptions, is the way to true high availability.
The active-active design of the NeoSapphire High Availability AFA ensures both nodes operate simultaneously, so if one node fails, the other immediately takes over, thereby limiting downtime to almost zero. In addition to continuous uptime and the need to make minimal recovery efforts, other advantages of the active-active architecture are its extremely low latency and excellent multi-path data access performance. These attributes give enterprises true always-on, always-available application services, including the ability to perform maintenance on applications without service disruption.
Recovery time objective (RTO) is the maximum targeted duration of time allowed between an unexpected failure or disaster and the resumption of normal operations and service levels. Unplanned disruptions can negatively affect business continuity and revenues. The NeoSapphire High-Availability AFA is based on a shared-nothing architecture, which leverages InfiniBand connectivity for the fastest data synchronization between data pools to achieve the lowest RTO.
The NeoSapphire High Availability AFA delivers world-class 99.9999% reliability. Based on a shared-nothing architecture and redundant hardware design, this model uses enterprise SSDs with a long lifespan suitable for write-intensive workload applications. These great features are powered by AccelStor’s award-winning software technology, FlexiRemap. This flash software optimizes I/O algorithms to maximize the SSD lifespan and data protection of each disk group. Every aspect of our HA model is designed to enhance reliability and availability. Its hardware architecture, software algorithms, and selection of SSDs combine to create a true high-availability all-flash array with essentially zero downtime.
The NeoSapphire High Availability AFA is engineered with an innovative design that places two complete systems in one storage appliance. That is, each storage node is a full system comprising both controller and data storage, so if a failure occurs at a single node, the NeoSapphire will still remain up and running. Reliability is also bolstered by the HA model’s active-active design, wherein both nodes operate simultaneously, so if one fails, the other immediately takes over. Plus, the nodes communicate with each other via a high-speed, low-latency InfiniBand FDR (fourteen data rate) interlink for dependable, synchronous replication. The status of the nodes is also monitored by a heartbeat private network connection between them, further ensuring system health.
All-flash arrays are often perceived as an expensive investment for IT infrastructures. In past years, IT procurement has left many organizations with oversubscribed and underutilized storage capacity. Looking seriously at the overall investment and ROI, the cutting-edge all-flash scenario is actually the best solution in terms of cost. TCO encompasses not only the initial purchase cost, but also maintenance costs and labor per terabyte. Compared with conventional HD arrays, all-flash arrays can reduce power usage by 76%, cooling costs by 63%, administrative fees by 48% and maintenance costs by 16%, in addition to saving 48% more space. What’s more, all-flash arrays dramatically improve enterprise application performance and decrease the costs of IT deployment.