Adding QoS features to an existing storage platform is just a temporary patch for a much larger problem. These features may solve one performance bottleneck under an isolated condition, but this approach falls apart at cloud scale as conditions multiply across a multi-tenant infrastructure. A true solution requires a purpose-built storage architecture that solves performance problems comprehensively, not individually.

With our benchmark storage architecture, SolidFire set the standard for delivering an exact level of performance and bandwidth within multi-tenant infrastructures. We built our storage architecture from the ground up to allow cloud providers to guarantee performance despite tenant activities or failure conditions, a solution made possible only through six core architectural elements:

	All-SSD Media Enables the delivery of consistent latency for every IO Without an all-SSD architecture, guaranteeing Quality of Service (QoS) is impossible. Tiered and hybrid disk solutions are all subject to unpredictable, rapid changes in IO latency that prevent predictable performance.
	True Scale-out Architecture Allows for linear, predictable performance gains as systems scale True Quality of Service (QoS) requires that storage systems are always aware of the total available performance. As systems scale, they must be capable of increasing both capacity and performance independently and in a predictably linear pattern.
	RAID-less Data Protection Ensures predictable performance in any failure condition To deliver guaranteed Quality of Service (QoS), systems must eliminate the performance variability caused by disk failures. Only RAID-less data protection delivers the predictable performance required for supporting firm performance SLAs.
	Balanced Load Distribution Eliminates hot spots that create unpredictable IO latency To deliver true storage QoS, systems must be capable of balancing both performance and capacity evenly across the entire system. Distribution of data based upon type – rather than location – eliminates hot spots caused by problematic application behavior.
	Fine-Grain QoS Control Completely eliminate noisy neighbors, and guarantee volume performance True storage Quality of Service (QoS) requires a robust model for configuring QoS settings for individual volumes. The system must also accommodate bursty workloads, changing performance requirements, different IO patterns, and the possibility of over-provisioning.
	Performance Virtualization Control performance independent of capacity and on demand Ensuring that performance resources are available on demand is essential to delivering Quality of Service. Achieving that capability requires being able to change performance levels on the fly without migrating data or affecting the performance of other applications.

Storage Technology

• Overview
• Performance/QoS
• Efficiency
• Management
• Scalability
• SolidFire Storage System
• SolidFire Element OS
• eScanner Tool
• OpenStack
• How To Buy

Related Resources

SolidFire Quality of Service (QoS) Benchmark Storage Architecture

Related Videos

Affordable all-SSD performance

Fine-grain QoS control for writing firm performance SLAs

Increase VM density to enable higher profits

Blog